Spring 2012 FRONTIERS OF ENGINEERING

The BRIDGE LINKING ENGINEERING AND SOCIETY

Frontiers in Additive Manufacturing: The Shape of Things to Come Hod Lipson Additive Manufacturing in Aerospace: Examples and Research Outlook Brett Lyons Large-Scale Visual Semantic Extraction Samy Bengio Challenges and Opportunities for Low- Carbon Buildings John Ochsendorf Using Information Technology to Transform the Green Building Market Christopher Pyke The Evolution of Brain-Computer Interfaces Eric C. Leuthardt Retinal Prosthetic Systems for the Treatment of Blindness James D. Weiland and Mark S. Humayun

The mission of the National Academy of Engineering is to advance the well-being of the nation by promoting a vibrant engineering profession and by marshalling the expertise and insights of eminent engineers to provide independent advice to the federal government on matters involving engineering and technology. The BRIDGE

NATIONAL ACADEMY OF ENGINEERING

Irwin M. Jacobs, Chair Charles M. Vest, President Maxine L. Savitz, Vice President Thomas F. Budinger, Home Secretary Venkatesh Narayanamurti, Foreign Secretary C.D. (Dan) Mote Jr., Treasurer

Editor in Chief: Ronald M. Latanision Managing Editor: Carol R. Arenberg Production Assistant: Penelope Gibbs The Bridge (ISSN 0737-6278) is published quarterly by the National Acad­ ­ emy of Engineering, 2101 Constitution Avenue, N.W., Washington, DC

20418. Periodicals postage paid at Washington, DC. Vol. 42, No. 1, Spring 2012 Postmaster: Send address changes to The Bridge, 2101 Constitution Avenue, N.W., Washington, DC 20418. Papers are presented in The Bridge on the basis of general interest and time- liness. They reflect the views of the authors and not necessarily the position of the National Academy of Engineering. The Bridge is printed on recycled paper. C © 2012 by the National Academy of Sciences. All rights reserved.

A complete copy of The Bridge is available in PDF format at http://www.nae.edu/TheBridge. Some of the articles in this issue are also available as HTML documents and may contain links to related sources of information, multimedia files, or other content. The Volume 42, Number 1 • Spring 2012 BRIDGE LINKING ENGINEERING AND SOCIETY

Editor’s Note 3 The Expanding Frontiers of Engineering Andrew Weiner

Features 5 Frontiers in Additive Manufacturing: The Shape of Things to Come Hod Lipson Additive manufacturing is the ultimate tool that may change human life in ways we can barely imagine. 13 Additive Manufacturing in Aerospace: Examples and Research Outlook Brett Lyons Additive manufacturing has the potential to revolutionize the production of aerospace and defense components. 20 Large-Scale Visual Semantic Extraction Samy Bengio Machine-learning algorithms for image annotation can enable the classification of huge amounts of different types of data. 26 Challenges and Opportunities for Low-Carbon Buildings John Ochsendorf Improving the construction and performance of buildings could lead to huge reductions in carbon emissions. 33 Using Information Technology to Transform the Green Building Market Christopher Pyke The green building community is using information technology to drive fundamental changes in the real estate market. 41 The Evolution of Brain-Computer Interfaces Eric C. Leuthardt Brain-computer interfaces create alternate communication channels for people with severe motor impairment. 51 Retinal Prosthetic Systems for the Treatment of Blindness James D. Weiland and Mark S. Humayun Researchers are working on implantable neurostimulators to treat debilitating neurological conditions, including blindness.

(continued on next page)

The BRIDGE

NAE News and Notes 58 Class of 2012 Elected 62 NAE Newsmakers 63 Changing the Conversation Is Now on Facebook 64 New Mirzayan Science and Technology Policy Graduate Fellows 65 The Asad, Taj, and Jamal Madni Matching Gift Challenges 66 A Message from NAE Vice President Maxine L. Savitz 68 National Academy of Engineering 2012 Private Contributions 78 Calendar of Meetings and Events 78 In Memoriam 78 NAE Annual Meeting, September 30–October 1, 2012

79 Publications of Interest

The National Academy of Sciences is a private, nonprofit, self- The Institute of Medicine was established in 1970 by the National perpetuating society of distin­ guished­ scholars engaged in scientific Acade­my­ of Sciences to secure the services of eminent members of and engineering research, dedicated to the furtherance of science and appropriate pro­fes­sions in the examination of policy matters pertaining technology and to their use for the general welfare. Upon the author- to the health of the public. The Institute acts under the responsibility ­ity of the charter granted to it by the Congress in 1863, the Academy given to the National Academy of Sciences by its congressional char- has a mandate that requires it to advise the federal gov­ernment­ on ter to be an adviser to the federal government and, upon its own scientific and technical matters. Dr. Ralph J. Cicerone is president of the initiative, to identify issues of medical care, research, and education. National Academy of Sciences. Dr. Harvey V. Fineberg is president of the Institute of Medicine.

The National Academy of Engineering was established in 1964, The National Research Council was organized by the National under the charter of the Nation­ al­ Academy of Sciences, as a parallel Academy of Scienc­ ­es in 1916 to associate the broad community of organization of outstand­ ing­ engineers. It is autonomous in its adminis- science and technology with the Academy’s purposes of fur­ther­ing tration and in the selection of its members, sharing with the National knowledge and advising the federal government. Func­tion­ing in Academy of Sciences the responsibility for advising the federal gov­- accordance with general policies determined by the Academy, the ernment. The National Academy of En­gineer­ ing­ also sponsors engi- Council has become the principal operating agency of both the neering programs aimed at meeting national needs, encourages edu- National Academy of Sciences and the National Academy of Engi- cation and research, and recognizes the superior achievements of neering in providing services to the government, the public, and the engineers. Charles M. Vest is president of the National Academy scientific and en­gi­neer­ing communities. The Council is administered of Engineering. jointly by both Academies and the Institute of Medicine. Dr. Ralph J. Cicerone and Charles M. Vest are chair­ and vice chair, respectively, of the National Research Council. www.national-academies.org FALL 2006 3 Editor’s Note

to build customized functional parts in small lot sizes to meet end-users’ specifications. Two presentations from this session are included in this issue of The Bridge. Hod Lipson, of Cornell Uni- versity, describes the evolution of additive manufactur- ing technologies as a series of milestones in achieving human control over physical matter (p. 5). His article covers additive manufacturing from its current, rela- tively mature technologies (programming of shape) to new frontiers in which the internal composition, struc- Andrew Weiner ture, and even the function of materials can be realized through programmable printing technologies. The Expanding Frontiers of In the second article, Brett Lyons, a researcher at Boeing, describes aerospace applications of additive Engineering manufacturing. Because this industry has an especially Every year the U.S. Frontiers of Engineering (FOE) demanding manufacturing context that blends low- Symposium brings together approximately 100 out- volume economics, acute weight sensitivity, and the standing young engineers, ages 30 to 45, to share ideas need for highly controlled materials and manufactur- and learn about cutting-edge research on a wide range ing processes to ensure high levels of repeatability and of engineering topics. A unique characteristic of the reliability, it has focused directly on transitioning addi- symposium series is that participants are competitively tive manufacturing techniques from the laboratory and selected from researchers working across the spectrum model shop to the factory floor (p. 13). of engineering disciplines in academia, industry, and The second FOE session, chaired by Aleksandar government. FOE provides these emerging engineer- Kuzmanovic of Northwestern University and Amarnag ing leaders with a unique opportunity to learn about Subramanya of Google, featured four presentations on exciting research in engineering areas other than their semantic processing. The emergence of the web as a tool own and to meet and network with promising engineers for sharing information has led to a massive increase in working in different areas. the size of potential datasets that often include different The seventeenth U.S. FOE Symposium was held types of data, such as texts, images, diagrams, music, and September 19–21, 2011, at Google, Inc., in Mountain others. Recent developments in distributed processing View, California. The meeting was organized into (e.g., cloud computing) have enabled large-scale inves- independent sessions with the following themes: addi- tigations of statistical machine-learning algorithms to tive manufacturing, semantic processing, engineering infer meaning from these datasets. Researchers hope sustainable buildings, and neuroprosthetics. Seven to use these algorithms to infer author sentiment from papers based on this year’s presentations are included texts, disambiguate references to real-world entities, in this issue of The Bridge. and perform a variety of other tasks important for The session on additive manufacturing, organized by web searches and other applications. In “Large-Scale Carolyn Seepersad of the University of Texas at Austin Visual Semantic Extraction,” Samy Bengio, of Google and Michael Siemer of Mydea Technologies Corpora- Research, describes an approach to automatic image tion, focused on layer-by-layer fabrication of complex annotation that can scale to very large, varied datasets parts directly from CAD files without part-specific of images and labels (p. 20). tooling. Additive manufacturing has many strategic The focus of the session on engineering sustainable advantages over traditional manufacturing processes: buildings, chaired by Annie Pearce of Virginia Tech the production of complex internal and external part and John Zhai of the University of Colorado, was on geometries that cannot be made in any other way; rapid the emerging integration and transformation of the iteration through design permutations; and the ability architecture/engineering/construction industry, which The 4 BRIDGE is increasingly building sustainable buildings that ben- viduals communicate and participate more in the world efit the social, economic, and natural environments. around them (p. 41). Modern design concepts for high-performance build- In the final article, “Retinal Prosthetic Systems for ings, coupled with new building materials and advanced the Treatment of Blindness,“ James Weiland and Mark mechanical and electrical systems, require a compre- Humayun, of USC, review the history of electrical hensive understanding of integrated building elements stimulation of the nervous system and highlight recent and systems, including the humans who design, operate, clinical advances in retinal implants for the treatment and occupy them. Two articles from this session are of blindness. Although much remains to be done, some included in this issue of The Bridge. significant advances have been made that can improve In “Challenges and Opportunities for Low-Carbon quality of life for some blind people. The authors are Buildings,” John Ochsendorf discusses cutting-edge optimistic that limited improvements in vision will be benchmarking for building performance and life-cycle- possible in the foreseeable future (p. 51). cost. He presents case studies of ultra low-carbon In addition to the talks, FOE symposia include lively buildings designed by his team at MIT, summarizes key Q&A sessions, panel discussions, and activities that challenges to the design and implementation of low- encourage personal discussions and networking. In carbon buildings in the United States, and describes the 2011, a series of five “lightning talks” by researchers consequent new opportunities for engineers, whose role from Google, which hosted the meeting, added to the is crucial to the success of this changing market (p. 26). excitement. The dinner speaker, a traditional highlight In the next article, “Using Information Technology of FOE programs, was Alfred Spector, Vice President to Transform the Green Building Market,” Christopher of Research and Special Initiatives at Google, Inc. His Pyke, of the U.S. Green Building Council, approaches remarks centered on the evolution of computer science, the subject from an industry perspective. He describes how it has influenced the world, and how it will help how location-based services and social networks are humankind manage the grand challenges that lie ahead. driving market transformation for sustainable building. It has been my great privilege to serve as chair of the Chris discusses how the development of new tools for Organizing Committee for the U.S. FOE symposia for identifying, comparing, and rewarding high-achieving the last three years. The interdisciplinary approach projects can drive continuous improvement (e.g., green- of FOE events and the diversity of participants have house gas emissions reductions, water conservation, and made these meetings truly stimulating and memorable. public health benefits) in the construction, operation, I know I will miss participating in them in the future. and life-cycle costs of green buildings (p. 33). In closing, I want to express my gratitude to Janet The final session, on neuroprosthetics, chaired by Hunziker, NAE senior program officer, Elizabeth Timothy Denison, Medtronic, and Justin Williams, Weitzmann, program associate, and Lance Davis, University of Wisconsin, Madison, focused on the NAE Executive Officer, for their contributions to the engineering of technologies that can interface with the planning and implementation of this unique series. I human body, both for stimulating the nervous system also want to thank the sponsors of the 2011 sympo- and for processing and acting on signals from the ner- sium—Google, Inc., The Grainger Foundation, the Air vous system. Force Office of Scientific Research, the Department of In “The Evolution of Brain-Computer Interfaces,” Defense ASDR&E Research Directorate-STEM Devel- Eric Leuthardt, of Washington University, discusses opment Office, the National Science Foundation, brain-computer interfaces that can acquire brain sig- Microsoft Research, and Cummins Inc. nals and translate them into machine commands that reflect the intentions of the user. Eric describes emerg- ing technologies that can help severely impaired indi- Additive manufacturing is the ultimate tool that may change human life in ways we can barely imagine.

Frontiers in Additive Manufacturing The Shape of Things to Come

Hod Lipson

Additive manufacturing technologies—machines that can automatically fabricate arbitrarily shaped parts, pixel by pixel, layer by layer, from almost any material—have evolved over the last three decades from a limited num- ber of expensive prototypes to widely available, small-scale commodity pro- duction tools. But this trend is only the tip of a technological iceberg, a burgeoning “second industrial revolution” that is expected to transform every Hod Lipson is associate profes- aspect of our lives. It is easy to predict that, like any other new manufactur- sor, School of Mechanical and ing technology, additive manufacturing will lead to more material options, Aerospace Engineering, Cornell better resolution, faster production, easier and more reliable operation, and lower costs. But it is less obvious where this technology will go next. We can University. look at the evolution of additive manufacturing technologies past, present, and future as a series of milestones in human control over physical matter.

Printing Forms: Programming the Shape of Matter The first milestone, which is maturing today, has been unprecedented control over the shape of objects. Machines can now fabricate objects of almost any material—from nylon to glass, from chocolate to titanium—and with any complex geometry. This capability is transforming not only engi- neering, but also many other fields, from biology to archaeology to education to the culinary arts. The 6 BRIDGE

a b c

d e f

FIGURE 1 3D printing is a universal technology that will impact every industry and discipline. (a) Complex cookie with interior letter in chocolate. (b) Woman’s shoe with elaborate titanium heel. (c) A printed robot that was designed automatically. (d) A bio-printed bovine meniscus printed with live cells according to a CD scan. (e) A printed clay copy of a 3,000 year old cuneiform tablet. (f) A 90 percent printed 3-gram flapping, hovering ornithopter. Sources: Figures 1a, 1c, 1d, 1e, and 1f, courtesy of Cornell University and http://creativemachines.cornell.edu. Figure 1b courtesy of WTWH Media.

Complexity for Free the marginal cost of added complexity with additive Perhaps the most dramatic impact of 3D printing manufacturing is near zero. is that it can manufacture objects without factor- Why should we care about the marginal cost of com- ing their complexity. Fabricating a solid block costs plexity? Lowering the cost of manufacturing complex almost the same as fabricating an oddly shaped object products is obviously a good thing, but the consequences with curved surfaces and notches. In either case, the have profound implications. Industrial revolutions are printer scans back and forth, depositing material one triggered when a fundamental cost associated with pro- layer at a time. The only difference is in the pattern duction drops to zero, essentially taking that factor out of material deposition. Just as printing a picture of a of the cost equation. circle takes no longer than printing a picture of the The industrial revolution of the nineteenth century map of the world, printing a mousetrap takes no lon- occurred when steam engines replaced horses and ger and requires the same skills and resources as print- waterwheels, dramatically reducing the cost of power ing a paperweight. to near zero. Not only were traditional power sources Regardless of how we measure it—in production replaced, but the types of work that machines could time, material weight, energy waste, or production plan- perform led to a cascade of innovation, such as rail- ning—the addition of features hardly changes cost. In roads and factory automation. Similarly, the Internet some cases, added complexity can even reduce cost. For reduced the cost of disseminating information to near example, printing a block with a hole in it is cheaper zero. As a consequence, it also expanded the range and faster than printing a solid block. Therefore, in and types of media that could be distributed—not just stark contrast to conventional manufacturing where online newspapers, but also wikis, blogs, and user- every additional hole, surface, protrusion, and corner generated content. requires more planning, takes longer to produce, and One could argue that additive manufacturing has consumes more energy and possibly more raw material, drastically reduced the cost of complexity, bringing SPRING 2012 7

it to near zero. Initially, this simply implies that new From Bio-printing to Food Printing technology will gradually replace older, more expensive Every discipline will be affected by the unprecedented ways of making things. But in the long term, the range control over the shape and composition of matter. Imag- and types of objects that can be manufactured will also ine the implications for any field that involves the design increase dramatically. and fabrication of physical objects—from mechanical engineering to art and architecture. But additive manu- Personalized Manufacturing facturing technology can also change fields with no In addition to vast new design possibilities, manu- immediate connection to engineering or manufacturing. facturing will be personalized. This trend has profound Take, for example, the emerging field of 3D printing for economic implications for how future products will be medical applications. designed and used, who will design them, and where they Custom-Shaped Prostheses. One of the earliest will be made (Lipson and Kurman, 2010). Most impor- applications of health-related 3D printing was the fab- tant, the opportunity for anyone to design and produce rication of custom-shaped, complex prosthetic limbs complex products without the need for the resources and devices, particularly to address the key challenge and skills of traditional manufacturing will democratize of the interface point with the body. For example, a innovation and unleash the long tail of human creativ- good match between the socket of a limb prosthesis and ity. Look online today, and you will see thousands of a patient’s bony prominences or muscular tissue can objects for sale ready to be printed on demand, from greatly improve both functionality and comfort. custom-shaped hearing aids to flapping, hovering micro When the aesthetic appearance of a prosthesis is a air vehicles (Figure 1f), to authentic-looking replicas of factor, the exterior can be custom shaped to blend with ancient cuneiform tablets (Figure 1e). its surroundings or to match symmetrical features. In addition, the ability to shape internal cavities, add Printing Composition: Shaping the Internal strengthening girders, and shave material off non-load- Structure of Materials bearing components can all improve the weight-to- The second milestone, which has just appeared on strength ratio of the device and make it both stronger the horizon, is controlling the composition of matter— and lighter. With sophisticated, multi-material fabrica- not just shaping external geometry, but also shaping tion, the mechanical performance of prosthetic devices the internal structure of materials with unprecedented could be tailored even further to improve elasticity, fidelity. Using multi-material additive manufacturing shock-absorbance, and energetic performance. technologies, we can make materials within materi- als, embed and weave multiple materials into complex patterns, and co-fabricate entangled components. For example, we can print hard and soft materials in pat- Every discipline will be terns that create materials with bizarre new structural behaviors, such as materials that expand laterally when affected by unprecedented pulled longitudinally. control over the shape and We are eliminating traditional limitations imposed by conventional manufacturing when each part had to composition of matter. be made of a single material. Instead, microstructure can now be specified with micron-scale precision. This will make it feasible for you or me to print a custom ten- Another type of shaped prosthesis is the hearing aid. nis racket to enhance our backhands or for a doctor to A major factor in the comfort and effectiveness of hear- produce a replacement spinal disc implant tailored for ing aids is how they fit in the aural canal. With 3D print- a specific patient. ing, the patient’s canal can be optically scanned, and a The number of possibilities is enormous, but so far few tailored soft prosthesis can be fabricated almost instantly. theories can predict the properties of these new materi- Custom-Shaped Implants. One-size-fits-all is not a als, and few designers are prepared to exploit the new good compromise when it comes to hip replacements. design space. Clearly, we will need new design tools to The idea of using 3D printing to fabricate custom-shaped augment human creativity. implants dates back to the earliest days of freeform The 8 BRIDGE fabrication. Today, custom-shaped titanium or platinum Drug-Screening Models. An exciting opportunity implants can be fabricated with exactly the right size for fabricating complex, multi-cell heterogeneous tis- and shape, either parametrically scaled to fit the patient sue arrangements in three dimensions is the fabrication or produced directly from CT scan geometry of the of models for drug screening. Traditionally, drugs and original bone or a symmetrical healthy bone. Because other treatments are tested in petri-dish-like environ- complexity does not increase cost, printed implants can ments that are meant to replicate the anticipated target not only match the standard bone shape, but can also environment in which the drug or treatment will be include various cavities and connection points to make used. However, these petri-dishes or test-tubes are rela- bonding with existing tissue more compatible, reliable, tively simple compared to the real environment, both and effective. in the range of cell types and their spatial distribution. This mismatch often necessitates the use of animal models and other sophisticated, high-throughput test- ing procedures. Printing with live cells makes With 3D bio-printing, however, 3D spatially hetero- it possible to fabricate geneous tissue models can be fabricated to more closely resemble the target application of a drug. For example, heterogeneous tissue implants. if cancer cells can be fabricated directly in the shape and distribution of a tumor, the effectiveness of treat- ments could be tested in vitro. Such experiments could Bio-printing. Instead of using titanium and other provide a more realistic prediction of drug effectiveness engineering materials, implants (and other constructs) and shorten the cycle of drug development. can be fabricated directly from biological materials. Surgical Planning. A less obvious, but equally Early experiments involved 3D printing of bio- important application of 3D printing is preparation compatible scaffolds, which were later infused with for surgery. Non-routine, complex surgical procedures live cells and incubated before implantation. The often involve manipulation of tissue and tools through live cells gradually replaced the scaffold, resulting in a an intertwined, unknown environment. Some opera- custom-shaped live tissue-engineered implant. tions also require the preparation and attachment of Printed scaffolds are still prevalent, but printing with permanent or temporary plates or other devices in biological cells directly, with no scaffold at all, is now the patient’s body. Just as it is easier to put a puzzle possible. In this case, cells are first immersed in a bio- together the second time around, surgeons can sub- compatible hydrogel ink—a bio-ink—and then printed stantially reduce the time and improve the reliability into their target form. The bio-ink has two key, con- of an operation if they practice it beforehand. This is tradictory properties that make it tricky to develop. On especially important for complex, non-routine proce- one hand, it must be fluid enough to be printed—to flow dures being attempted for the first time. through the nozzle without damaging cells that are sub- For example, the veterinary school often asks us to jected to severe shear forces as they come out of the print- bio-print a 3D set of shattered or deformed bones from head. On the other hand, bio-ink must be stiff enough a CT scan of an injured animal that needs emergency to hold its shape after printing to prevent the material surgery. With the printed bones, surgeons can practice from oozing into a shapeless mass. To address this double the operation—learning to identify fragments, optimize challenge in fabricating cartilage implants in the shape reconstruction, and prepare plates and jigs in advance. of a meniscus directly from CT data (Figure 1d), we used Informally, surgeons have reported that surgery time a variety of chemical and optical cross-linking agents. can be reduced by half and that the quality of care is Unlike scaffold-infusion techniques, printing with greatly improved. live cells directly makes it possible to fabricate het- Surgical Training. Surgeons in training often have erogeneous tissue implants. Imagine the fabrication access to state-of-the-art surgical equipment and tools, of a form as complex as a spinal disk or a heart valve, but they rarely have a chance to practice on realistic which involves multiple cell types in a complex spatial cases. For example, a surgeon in training is unlikely to arrangement that is critical to its proper functioning be able to practice removing a brain tumor on an animal (Cohen et al., 2006). model, a human cadaver with this condition, or a syn- SPRING 2012 9

thetic training model. With 3D printing, however, rel- and its composition, we will also be able to program evant, typical cases from real patients can be recorded materials to function in arbitrary ways—to sense and and reproduced on demand for practice. react, to compute and behave. This last step entails With some modeling, it is even possible to combine a blurring of the line between material and code cases and adjust their severity on demand to challenge (Gershenfeld, 2005), leading to what is essentially surgeons-in-training at the appropriate level for the programmable matter, including everything from con- best learning experience. With multi-material bio- trolling the mechanical functionality of an object to printing, training models could be fabricated with controlling how it processes information and energy. biological materials to provide realistic experience When we reach this milestone, we will be able to that also provides the feel and responsiveness of real print virtually anything—from a cell phone to a robot wet tissue. that walks out of the printer, batteries included. But Custom Medications. An excellent example of an the robot will not look at all like today’s robots, because application of 3D printing is the fabrication of medical it will not be limited by the constraints imposed by pills on demand. A growing challenge for both doctors conventional manufacturing, and it won’t be designed and patients is the administration of multiple medi- directly by humans (Figure 1c). The ability to manu- cations simultaneously. Instead of keeping track of a facture arbitrary active systems comprising both passive dozen pills daily, a printer can now fabricate a single, and active substructures will have opened the door to a custom-made pill for each patient. The pill could have new space for designs and a new paradigm of engineer- an identification marking, which would eliminate the ing—one that is not unlike biology. confusion and uncertainty associated with conven­ tional delivery methods. The New Computer-Aided Design Food Printing. Food printing is to 3D printing what For the last four decades, computer-aided design video gaming is to computers. Based on the same tech- (CAD) tools have played a critical role in product nology as bio-printing, food printing is the fabrication of design. But the role and format of CAD tools have edible items from raw edible inks. “Edible inks” include remained relatively unchanged. Although user inter- chocolate and peanut butter, cookie dough and frosting, faces have improved, geometric manipulations have as well as organic pesto and locally made goat cheese. It become faster and more reliable, and graphics have doesn’t really matter: Download the recipe, load in the become three-dimensional and photo-realistic, con- frozen food cartridges, and hit print. ceptually CAD software remains a passive 3D drawing The recipe dictates which material goes where and board that records intentions but offers little insight or which in-line-cooking procedure is applied during ideas of its own. deposition and after. We have printed foods with all of the materials mentioned above and more, creating chocolate confections with frosted decorations and vanilla cookies with chocolate vertical text lettering The last step will blur the inside (Figure 1a). Imagine a cookie printer with sliders line between material and to adjust crispiness, flavor, color, and texture. Imagina- tion is the only limiting factor. code leading to essentially Although some may consider printed foods the epit- ome of processed foods, others are fascinated by the programmable matter. opportunities for almost unlimited innovation. You can download and share recipes, tailor variants as you wish, and produce freshly made dishes to your exact liking. As 3D printing technologies become more abundant, traditional barriers of resources and skills are all but van- Printing Function: Programming the Behavior ishing. The limit is now our imaginations, but, unfor- of Active Materials tunately, our imaginations are limited. Over and over The third and final milestone, of which we are just again, I have seen students faced with the blank page of beginning to see early signs, is control of behavior. At CAD and the unlimited capability of a 3D printer design this point, we will not only control the shape of matter rectangular objects with a few linear notches. The 10 BRIDGE

To a large extent, this cultural blindness is the result The spherical surface might be more easily described of years of observing mass-produced objects subject to using algorithmic rather than descriptive geometry, a traditional manufacturing constraints. However, it also procedural construction process rather than a target reflects design thinking imposed by conventional CAD geometry (Pasko et al., 2011)—much the way a biolo- tools and the lack of new tools that can take advantage gist describes phenotypes using a developmental pro- of the vast new design space offered by 3D printing capa- cess or a software engineer describes the appearance of bilities. Clearly, the classical CAD paradigm will be a dynamic web page. dominant for the foreseeable future, but new paradigms In fact, there are many more crazy shapes than regu- for design tools are beginning to emerge. lar shapes (Figure 2a), but they are currently difficult Function Representations. As our ability to con- to explore. If designers were unencumbered by tra- trol the shape, composition, and behavior of materials ditional manufacturing constraints, a growing num- advances, it becomes more appropriate to think about ber of them would be anxious to explore these kinds geometry and material specification as programming of geometries. rather than drawing. For example, say you want to Matter Compilers. An alternative approach to fabricate a spherical surface with periodic patterned design is to specify what the design should accom- protrusions (Figure 2b). Fabricating such an object plish rather than what it should look like and then using traditional manufacturing techniques would be a let the machine compile the design to meet your complicated, expensive nightmare. As a result, such a specifications. Some products, especially products capability is buried far down the advanced CAD options with purely functional roles, would fit well into such a menu, if it can be found at all. design process.

a b

c

FIGURE 2. The New CAD. Beyond geometric modeling, products are described automatically by interactive evolution, through procedural construction algorithms, or by compiling high-level requirements and constraints. Sources: Figure 2a courtesy of Cornell University and http://endlessforms.com. Figure 2b: Pasko et al., 2011. Courtesy of Turlif Vilbrandt. Figure 2c designed by Jon Hiller. Reprinted courtesy of Cornell University and http://creativemachines.cornell.edu. SPRING 2012 11

Consider, for example, designing a supporting bracket. Even though a toothbrush seems like a simple product, We know the geometric constraints of the load and sup- it takes years of experience and know-how to design a port contact points, the weight to be carried, and the successful one. Nevertheless, with 3D printers, people material properties. Specify those requirements, hit the with almost no experience and little patience to learn design button, and watch the optimal design emerge are likely to want to design their own. automatically. I guarantee that it will not be a block The solution may be simple CAD applications for a with rectangular notches and holes, but an organic-look- specific product that encapsulate all relevant knowl- ing optimal structure with beautifully shaped cavities, a edge, yet provide just the right level of flexibility to the design that would take a human designer years to come user. Such FabApps (a term coined by Daniel Cohen up with manually. and Jeffrey Lipton), similar to the iPhone apps you can Now imagine that the bracket can’t be installed download for 99 cents, could guide you through the because of a protruding pipe. No problem—just add the design process and make you look like a pro. A tooth- pipe protrusion constraint and recompile (Figure 2c). brush app could ask for the dimensions of your hand and Programming by specifying target behavior and con- mouth, process pictures of your face and palm, walk you straints, then compiling it into functional geometry, can through 50 different options, ask 20 more questions, and address complex requirements, while simultaneously then produce a toothbrush that fits your unique needs exploiting the new manufacturing capabilities afforded perfectly and is guaranteed to be a success. by multi-material 3D printing as they become available. In a way, the designer becomes the customer, and CAD Conclusion software becomes the designer. Tool-making is one of the characteristics that distin- Interactive Evolution. What happens when a desired guishes modern humans from their evolutionary ances- design goal can’t be described quantitatively, such as an tors. Additive manufacturing, the ultimate tool, has the object that has an aesthetic component that defies quan- potential to change human culture forever in ways we tification? Imagine, for example, designing a perfume can hardly anticipate. bottle. Some aspects of the design, such as volume and size, can be specified quantitatively, but others, such as References the look and feel, are more difficult to describe. More- Clune, J., and H. Lipson. 2011. Evolving Three-Dimensional over, people skilled in the art of assessing perfume bottles Objects with a Generative Encoding Inspired by Develop- may have no inclination to dabble in “CAD speak.” mental Biology. Proceedings of the European Conference Instead, a new type of CAD software can display a on Artificial Life. See http://EndlessForms.com. range of initial design concepts and allow an expert Cohen, D.L., E. Malone, H. Lipson, and L. Bonassar. 2006. to indicate the ones he or she likes best or dislikes 3D direct printing of heterogeneous tissue implants. Tissue least. With this information, the machine can auto- Engineering 12(5): 1325–1335. matically infer a sense of the aesthetic the designer Gershenfeld, N. 2005. FAB: The Coming Revolution on has in mind and generate new solutions that meet Your Desktop—From Personal Computers to Personal Fab- the quantitative requirements but are also closer to rication. New York: Basic Books. the designer’s aesthetic preferences. By repeating this Hiller, J., and H. Lipson. 2009. Design Automation for Multi- process through a sequence of solutions and selections, Material Printing. Solid Freeform Fabrication Sympo- a designer with little CAD expertise might design a sium (SFF’09), August 3–5, 2009, Austin, Tex. Available complex perfume bottle. online at http://creativemachines.cornell.edu/sites/default/files/ In addition, the designer might come across new ideas SFF09_Hiller2.pdf. and be provoked into exploring corners of the design Knapp, M., R. Wolff, and H. Lipson. 2008. Developing space where no one else has been. Using this same pro- Printable Content: A Repository for Printable Teaching cess, visitors to the EndlessForms website (Clune et al., Models. Proceedings of the 19th Annual Solid Free- 2011) have collaboratively designed a variety of objects, form Fabrication Symposium, Austin, Tex., August 2008. from furniture to faces and from bottles to butterflies. Available online at http://creativemachines.cornell.edu/ FabApps. What if I want to design a toothbrush to papers/SFF08_Knapp.pdf. print on my 3D printer today? It is unlikely that I will Lipson, H., and M. Kurman. 2010. Factory@Home: be able to design a good, ergonomic, safe toothbrush. The Emerging Economy of Personal Fabrication. The 12 BRIDGE

Washington, D.C.: Office of Science & Technology Pol- the Range of DMLS of Titanium, Make Parts Fast. Avail- icy. Available online at http://web.mae.cornell.edu/lipson/ able online at http://www.makepartsfast.com/2009/12/828/ FactoryAtHome.pdf. explore-the-range-of-dmls-of-titanium/. Lipson, H., and J.B. Pollack. 2000. Automatic design and Pasko, A., O. Fryazinov, T. Vilbrandt, P-A. Fayolle, and V. Manufacture of Robotic Lifeforms. Nature 406(6799): Adzhiev. 2011. Procedural function-based modelling 974–978. Available online at http://creativemachines. of volumetric microstructures. Graphical Models 73(5): cornell.edu/papers/Nature00_Lipson.pdf. 165–181 Lipton, J.I., D. Arnold, F. Nigl, N. Lopez, D.L. Cohen, N. Richter, C., and H. Lipson. 2010. Untethered Hovering Noren, and H. Lipson. 2010. Multi-Material Food Print- Flapping Flight of a 3D-Printed Mechanical Insect. Pp. ing with Complex Internal Structure Suitable for Conven- 797–803 in 12th International Conference on Artificial tional Post-Processing. 21st Solid Freeform Fabrication Life (Alife XII), Odense, Denmark, August 2010. Avail- Symposium (SFF’10), Austin, Tex. able online at http://creativemachines.cornell.edu/sites/ Luft, K. Titanium Footware, See L. Carrabine, (2009) Explore default/files/Alife10_Richter.pdf. Additive manufacturing has the potential to revolutionize the production of aerospace and defense components.

Additive Manufacturing in Aerospace Examples and Research Outlook

Brett Lyons

The advantages of additive manufacturing are now widely recognized, even in the general media, and are predicted to revolutionize manufacturing processes for many industries (Economist, 2011). For aerospace, complex additive manufacturing processes must be developed to meet the industry’s stringent requirements and to ensure that products can achieve the robust performance levels established by traditional manufacturing methods. This Brett Lyons is a materials and pro- article provides an overview of additive manufacturing technologies being cess engineer, Boeing Research and used for the production of aircraft components and examples showing the Technology, The Boeing Company. direction of ongoing research and related developments.

Aerospace Requirements for Manufacturing To meet the stringent conditions necessary to ensure safety in air travel, aerospace manufacturers must satisfy a long list of complex requirements for even the simplest part. The consistent production of parts with identi- cal, well-understood properties requires that both materials and production processes be understood to a very high level. This complicated manufactur- ing context, which blends low-volume economics, acute weight sensitivity (weight is often the deciding factor in choosing materials and manufacturing processes for aerospace and defense components), and the need for highly controlled materials and manufacturing processes, has led Boeing to focus on transitioning additive manufacturing techniques from the laboratory and model shop to the factory floor. The 14 BRIDGE

Requirements for commercial aircraft parts are based facturing processes, which can form finished parts on U.S. Federal Aviation Regulations, which must be without intermediate tooling, thus eliminating the asso- met before a Type Certification can be issued for a given ciated costs and delays, are extremely attractive to the aircraft series; this certification is required for service with industry (Ruffo et al., 2006). any airline (FAA, 2011). The regulations are extensive and detailed, but the single most pertinent language in Selective Laser Sintering the context of additive manufacturing can be found in One of these processes, selective laser sintering (SLS), Title 14, Section 25, Subpart D, Subsection 25.605: begins with a computer-generated, three-dimensional design of a given part. The part is then digitally seg- The suitability and durability of materials used for parts, the failure of which could adversely affect safety, must mented into very thin layers that are selectively solidi- (a) Be established on the basis of experience or tests; fied in the machine, layer by layer. Essentially, the (b) Conform to approved specifications (such as indus- machine “grows” the part. With this capability, SLS can try or military specifications, or Technical Standard produce components with complex designs that would Orders) that ensure their having the strength and other be extremely complicated and expensive to produce by properties assumed in the design data; and (c) Take into other processes. account the effects of environmental conditions, such as SLS can produce economical thermoplastic parts that temperature and humidity, expected in service. are lightweight, nonporous, thin-walled, and highly This brief, but clear requirement is just one of many complex geometrically. Since the first implementations that have not only contributed to the incredible safety of SLS on Boeing aircraft, this process has been adopted record of commercial air transportation, but have also by a large number of programs, both military (Wooten, provided the impetus for studying new fabrication 2006) and commercial (Lyons et al., 2009), as well as by methods. Every aerospace and defense manufacturer producers of unmanned aerial vehicles. either has internal specifications or looks to established SLS produces lightweight, highly integrated sys- standards organizations for data that can support the tems and payload components (Figure 1) and, at the accurate design of components from a given mate- same time, eliminates non-recurring tooling costs and rial, based on minimum allowable performance levels. provides for life-cycle production flexibility. Because Factors that must be considered for the material per- weight is often a critical factor, SLS, which can produce formance of even the simplest components include spe- very thin walls and complex designs, is also attractive cific strengths, fatigue resistance, creep resistance, use for replacing parts that are typically produced using temperature, survival temperature, several tests of flam- established processes, such as rotational, injection, or mability, smoke release and toxicity, electric conductiv- polymer matrix composite molding. ity, multiple chemical sensitivities, radiation sensitivity, Drafting commercially efficient specifications in the appearance, processing sustainability, and cost. unique context of aerospace manufacturing requires Despite increased rates of production, the aerospace that a company develop in-depth knowledge of mate- industry must still produce many parts in very small rials and processes. For example, to take advantage quantities (Boeing Company, 2011). Additive manu- of SLS, one must have a firm understanding of the extreme four-dimensional energy input gradients dur- ing processing. Typical SLS machines use 75-Watt car- bon dioxide (CO2) lasers that have a 500-µm spot size. The laser spot moves at speeds of up to 10 meters per second over layers of nylon powder only 100 µm thick; each layer is com- pleted in approximately Figure 1. Photograph and illustration of laser-sintered air ducts. 60 seconds. A thorough SPRING 2012 15

Figure 2. Multi-zone heating (left), SLS part bed (center), and the same parts seen via infrared thermography during laser scanning (right). description of both the SLS process and efforts to simu- New Performance Challenges late details of the energies present can be found in the As the number of potential additive manufacturing literature (Franco et al., 2010). applications in aerospace has increased, three new per- formance challenges have arisen for SLS polymer mate- Examples of Aerospace-Driven Research rials. Operating requirements for the F-35 and B787 Controlling Temperature in the Part-Building Platform have challenged researchers to develop materials that To build parts with repeatable mechanical properties (1) can operate at higher temperatures, (2) have signifi- and dimensional control, the temperature distribution cantly better flame resistance, and (3) offer adjustable across the part-building platform must be as even as pos- electrical conductivity (Shinbara et al., 2010). sible. To accomplish this and reduce scrap rates, Boe- To add to these challenges, the new physical perfor- ing and its partners at the University of Louisville and mance targets must be met while maintaining as many Integra Services International (Belton, Texas) devel- of the attributes already established for SLS polyamide oped a patented method for zonal control of the tem- materials as possible. These attributes include mechan- perature of the part bed in SLS equipment (Huskamp, ical toughness, resistance to chemical attack, resistance 2009). Multi-zone, near-infrared (IR) wavelength to ultraviolet radiation, dimensional fidelity, and viable heating elements (Figure 2) provide the rapid response economics. and spatial resolution necessary to maintain a steady, Historically, when a new technology is enabled and even temperature. This invention, when paired with nears transition to useful service, researchers in many real-time IR imaging, has improved thermal control to industrialized nations tend to work on the same techni- a more advanced level than is required for most other cal problems simultaneously. In this case, researchers thermoplastic processing methods. in the United States, Germany, Japan, and the United Kingdom have conducted in-depth investigations into The Development of Flame-Retardant Polyamides high-performance polymers for SLS (Hesse et al., 2007; Another aerospace-driven need that researchers have Kemmish, 2010). addressed is for flame-retardant polyamides. Consider- If such a material were developed successfully, it ing that many polymers are derived from fossil fuel- could have a wide range of applications in and beyond based hydrocarbon feedstocks, the requirement that a aerospace. Two notable non-aerospace applications for related chemistry be self-extinguishing when exposed new high-performance, SLS-produced polymers are for to flame poses a serious challenge. Nevertheless, flame- medical implants and devices (Schmidt et al., 2007) retardance is necessary to a greater or lesser degree for and the production of low-volume automotive parts. all polymer materials used in the interiors of commercial aircraft. To take advantage of the weight and manufac- The Focus in Aerospace Research turing benefits of SLS on its commercial aircraft, Boeing Although many high-performance polymers are collaborated with its suppliers to develop the first laser- attractive from a cost perspective, the high cost of test- processed material that could pass the required flamma- ing for aerospace applications makes focusing on mul- bility tests (Booth et al., 2010). tiple materials simultaneously cost prohibitive. Unlike The 16 BRIDGE injection molding and other methods of polymer pro- information (Kemmish, 2010; Kohan, 1995). By com- cessing that rely on both heat and pressure to form a paring the bulk thermal properties of these two material part against the surface of a mold, additive manufac- families, one can begin to understand how differently turing processes rely primarily on the input of thermal they will behave in the SLS process. energy. Thus, viable materials for additive processes One important difference is the amount of energy must have very specific viscosity and other properties required to heat the material. To process a polyamide to be processed successfully. powder, the lower melt temperature combined with the lower specific heat (the amount of energy required to Polyaryletherketone (PAEK) Materials vs. Polyamides heat a given mass of material one degree Kelvin) indi- Because of their known performance, the polyaryl­ cates that the effort required to achieve a given viscosity etherketone (PAEK) family of materials is considered with heat input is much lower for a polyamide than for the lowest risk option for current development for aero- a PAEK. The PAEK must be heated to twice the tem- space applications. The PAEK family includes different perature just to approach melt and requires almost twice chemistries (e.g., polyetherketone, polyetherether­ the energy per degree of heating. This is further compli- ketone, and polyetherketoneketone). The choice of cated in SLS processing because the transient heating a PAEK as the next material to be developed is based requirements of each layer must not change drastically. on inherent factors, including very good flammability A second difference is the ratio of specific heat to the and chemical resistance, low moisture sorption, good heat of fusion (the energy flux exhibited in the transi- mechanical performance, good resistance to creep and tion from solid to liquid, and vice versa). In the SLS fatigue, compatibility with several methods of steriliza- process, the polymer powder is heated in stages from tion, and numerous material grades and suppliers to ambient conditions to very near the melt point. Poly- choose from. In light of the comparatively small size of amide material has a lower ratio of specific heat to heat the SLS market and the cost of developing new polymer of fusion than PAEK materials (1.26/226 compared to chemistries, raw materials for SLS are typically selected 2.20/130). This is important because, despite our best from commercial off-the-shelf grades of PAEKs, which efforts, there is still some gradient of energy input and have been designed for coatings, films, rotational mold- temperature across the building area at any given time. ing, and other applications. If a material very gradually transitions into melt, as fully At the time of this writing, the development of a amorphous polymers do, it can be difficult to feed the viable PAEK-SLS material is an area of competitive material smoothly onto the machine’s part-building area. industrial research, so specific information from any one The ratio of specific heat to heat of fusion gives an party is generally not published. However, a compari- indication of how easily a given material can be heated son of established, well-understood polyamides and the to near the melt point, across the whole part bed, with- PAEK family of materials shows the problem space for out fusing particles together. The closer to the melt engineers working in this field. point the material is when fed into the machine, the Table 1 shows comparative thermal properties of lower the energy input requirements on the laser for lower temperature polyamides and PAEK materials, as heat input to transition the material into the melt found in the literature and in manufacturer-published region. The lower the requirement on the laser for

TABLE 1 Comparison of the Thermal Properties of Polyamides and PAEK Family Materials

Material Melt Glass Transition Specific Heat of Fusion Thermal Thermal Specific Temperature Temperature Heat (100 percent Conductivity Expansion Gravity (g/cc) (°C) (°C) (J/g-K) crystalline) (J/g) (W/m-K) (ppm/Tg °C) (crystalline)

Polyamide 180–186 42–55 1.26 226 0.19 85 1.03

PAEK family 300–375 145–165 2.20 130 0.26 60 1.30 materials

Note: J/g-K = joules per gram-Kelvin. J/g = joules per gram. W/m-K = watts per meter-Kelvin. Ppm/Tg °C = parts per million per transition in degrees centigrade. SPRING 2012 17

energy input, the lower the risk of polymer degradation. This is because in the CO2 laser spot there is a roughly Gaussian distribution of energy, the peak of which can cause degradation. With a given energy input requirement on the laser (per the comparison above), the layer can be drawn with faster or slower laser scan speeds. The scan FIGURE 3 Examples of parts made of PAEK materials processed via SLS. speed affects the overall per-layer time, which, in addition to the proportion of some parts. Although this has provided a good source of preheat to laser energy required, results in a variable revenue to support new companies, it has also impeded temperature distribution and cooling rate for a cross new applications by making the development of new section of the part, per given layer. materials difficult for all but the largest users and mate- If too much time passes between the start and stop of rial companies. a given layer and the energy demand on the laser is too Dependence on the sale of materials and parts, as high, some sections of that layer will cool faster than well as nonproductive patent litigation, have also dis- others and the layer will not shrink uniformly. Dimen- tracted machine manufacturers from improving upon sional distortion can result if the cooling gradient is too their equipment with an eye toward higher volume, high relative to the recrystallization temperature, ther- economical, industrial manufacturing. Equipment mal conductivity, coefficient of thermal expansion, and manufacturers (e.g., Toshiba, Haas Automation, MAG, a host of other factors. Husky, and Arburg) cannot rely on the sales of materi- A thorough description of the interaction, just for als or parts to bolster their businesses. Thus for the the properties shown in Table 1, is beyond the scope of additive manufacturing industry to grow, it might look this paper, but these examples provide a window into to the business models and history of such machine the problems currently being addressed by researchers. manufacturers for reference. Thankfully, despite the complexities, several research teams are reporting success with the processing of PAEK Manufacturing Metal Parts materials via SLS (Figure 3). The use of metals in additive manufacturing for aero- space is as complex and exciting as the development Moving Forward of polymers. So far, engine manufacturers have shown Transitioning from the Laboratory to the Factory leadership in the direct manufacturing of metal parts, Numerous research frontiers lie beyond the develop- which is a highly dynamic field. However, leverag- ment of higher performing polymers. For one thing, ing the capabilities of additive manufacturing to cre- additive manufacturing equipment must transition from ate components with highly complex shapes also lends comparable low-reliability laboratory-grade equipment itself to tool production, for both metal and composite to hardened, cost-effective, high-temperature, industrial- components. In fact, new tooling-focused machines grade machines. The additive manufacturing industry and materials are being actively developed to leverage can look to its predecessors in injection molding and the process benefits of additive manufacturing for faster computer numerical-control machining for examples of production of metal castings (Halloran et al., 2010) and establishing new manufacturing technology and devel- long-fiber-reinforced composites (Wallen et al., 2011). oping a supporting business case. Analysis of Complex Geometries So far, in the infancy phase, the unique material requirements of additive manufacturing processing have Independent of material and processing conditions, limited machine manufacturers to selling materials and the analysis of complex geometries that can be built The 18 BRIDGE

FIGURE 4 Two complex trusses suggest the difficultly of predictive analysis. only with additive methods is another important area of Fire Retardant Materials. U.S. Patent Application active research. Even with material test data, the pre- 2010/0255327, October 7. dictive behavior of the types of structures additive man- Economist. 2011. Print me a Stradivarius: How a new manu- ufacturing can build, such as trussed airfoils (Figure 4), facturing technology will change the world. Economist is difficult to analyze. This field of study is focusing on (February 12): 11. developing new software tools for the generation and Engelbrecht, S., L. Folgar, D.W. Rosen, G. Schulberger, and predictive analysis of complex structures, such as three- J. Williams. �������������������������������������������2009. Cellular Structures for Optimal Per- dimensional trusses (Engelbrecht et al., 2009). formance. Available online at http://utwired.utexas.edu/lff/ symposium/proceedingsArchive/pubs/Manuscripts/2009/2009- Conclusion 73-Rosen.pdf. The descriptions of current research, along with FAA (Federal Aviation Administration). 2011. FAA Regu- examples, such as Boeing’s use of SLS for producing lations. Available online at http://www.faa.gov/regulations_ parts for commercial and military aircraft, show that policies/faa_regulations/. Retrieved July 18, 2011 the aerospace industry has the opportunity to lead, Franco A., M. Lanzetta, and L. Romoli. 2010. Experimental and the responsibility to contribute to, this revolu- analysis of selective laser sintering of polyamide powder: an tionary field of manufacturing technology. Technolo- energy perspective. Journal of Cleaner Production 18(16– gies that provide this level of manufacturing flexibility 17): 1722–1730. will greatly contribute to meeting burgeoning global Halloran, J.W., V. Tomeckova, S.P. Gentry, S. Das, P. Cili- demands for energy, transportation, defense, and medi- no, D. Yuan, R. Guo, A. Rudraraju, T. Wu, W. Baker, D. cal technologies here on earth and may even prove Legdzina, D. Wolski, W. Zimbeck, and D. Long. 2010. to be enabling technologies for exploring horizons Photopolymerization of powder suspensions for shap- beyond our planet. ing ceramics. Journal of the European Ceramic Society 31(14): 2613–2619. References Hesse, P., T. Paul, and R. Weiss, 2007. Powder��������������������� for Rapid Pro- Boeing Company. 2011. Boeing to increase 737 Production totyping and Associated Production Method. UP Patent Rate. Boeing Press Release. Available online at http://boe- Application 2007/0267766, November 22, 2007. ing.mediaroom.com/index.php?s=43&item=1426. Huskamp, C.S. 2009. Methods and Systems for Controlling Booth, R., B.C. Thornton, D.L. Vanelli, and M.L. and Adjusting Heat Distribution over a Part Bed. U.S. Gardiner. 2010. Methods and Systems for Fabricating Patent 7,515,986. Date issued April 7, 2009. SPRING 2012 19

Kemmish D. 2010. Update on the Technology and Applica- Shinbara, T. Jr. 2011. High Temperature Laser Sintered tions of Polyaryletherketones. Shawbury, U.K.: iSmithers Components. Presented at the Society of Manufacturing Rapra Publishing. Engineer’s AeroDef Manufacturing 2010 Exposition and Kohan, M.I. 1995. Nylon Plastics Handbook. Cincinnati, Conference, Anaheim, Calif., April 7, 2011. Ohio: Hanser Gardner Publications. Wallen, M., J. Rossfeldt, C. Aune, and Z. Wing. Washout Lyons, B., E. Deck, and A. Bartel. Commercial Aircraft Tooling for Hollow Composites: Toward Rapid, Low Cost Applications for Laser Sintered Polyamides. SAE Tech- Production Technology. Paper presented at Society of nical Paper 2009-01-3266. 2009 Warrendale, Pa.: SAE Manufacturing Engineers Composite Manufacturing Con- International. ference and Exhibits, April 12–14, 2011, Dayton, Ohio. Ruffo, M., C.J. Tuck, and R.J.M. Hague. 2006. Cost Estima- Wooten, J. 2006. Aeronautical Case Studies Using Rapid tion for Rapid Manufacturing—Laser Sintering Production Manufacture. Chapter 15 in Rapid Manufacturing: An for Low to Medium Volumes. Proceedings of the Institu- Industrial Revolution for the Digital Age, edited by N. tion of Mechanical Engineers, Part B: Journal of Engineer- Hopkinson, R. Hague, and P. Dickens. Chichester, U.K.: ing Manufacture 220(9): 1417–1427. John Wiley & Sons Ltd. Schmidt, M., D. Pohle, and T. Rechtenwald. 2007. Selective Laser Sintering of PEEK. CIRP Annals - Manufacturing Technology 56(1): 205–208. Machine-learning algorithms for image annotation can enable the classification of huge amounts of different types of data.

Large-Scale Visual Semantic Extraction

Samy Bengio1

The emergence of the web as a tool for sharing information has led to a massive increase in the size of potential datasets. Millions of images on web pages have tens of thousands of possible meanings that can be associated with labels, or annotations, in the form of HTML tags. These tags are used by webpage authors to describe the content of sections of web pages and by search engines or other web pages to link to them. Thus, HTML tags can be Samy Bengio is a research scien- conveniently collected by querying search engines, such as www.flickr.com tist in machine learning, Google (Torralba et al., 2008), or human-curated labels, such as www.image-net.org Research. (Deng et al., 2009). Clearly, we need machine-learning algorithms for image annotation that can learn from and annotate data on this large scale. The algorithms must include (1) scalable training and testing times and (2) scalable memory usage. The ideal algorithm would be fast and would fit on a laptop, at least at annotation time. Although many models have been proposed and tested on small datasets, it is not clear if any of them satisfies these requirements. In the first part of this article, we describe models that can learn to repre- sent images and annotations jointly in a low-dimensional embedding space. The low dimension simultaneously implies fast computations for ranking annotations and small memory usage. To ensure good performance for such a

1 This article is based on Weston et al., 2010, and Bengio et al., 2010. SPRING 2012 21

model, we propose estimating its parameters by learning One common approach includes (1) learning a dic- to rank annotations and optimize the process so that at tionary of common features from a large repository of least the top ranked annotations are relevant for a given images; (2) using an unsupervised clustering approach, image (this process is often called optimizing precision- such as K-Means; and (3) representing the image as a at-top-k). histogram of the number of times each feature of the In the second part of the article, we propose a novel dictionary is present in the image. algorithm for shortening testing time in multi-class Assuming the image representation is a d-dimensional classification tasks in which the number of classes (or real-valued vector, we denote it by x ∈ℜd. Further- labels) is so large that even a linear algorithm can more, let us assume our annotation task can consider become computationally infeasible. Our proposal is for up to Y possible text labels (or annotations). We can an algorithm for learning a tree structure of the labels in then represent each annotation as a unique integer the joint embedding space. We believe that by optimiz- iY∈ϒ = {}1,, index into the dictionary of all pos- ing the overall tree loss (i.e., the loss one would incur by sible annotations. using the tree to retrieve the top-k labels for any given Finally, consider an arbitrary D-dimensional embed- image), such an algorithm would be more accurate than ding (or target) space into which we want to map both existing tree-labeling methods. images and labels. Our goal is then to learn to trans- form images, represented in their original image-feature Joint Embedding of Images and Labels spacex ∈ℜ d into a new representation in the joint space D dD Imagine that one could transform different types of ℜΦ::I (x) ℜ→ℜ while jointly learning a map- objects—images, videos, music tracks, or texts—into a ping for annotations (i.e., a position in the new joint common representation (such as a vector of real num- space for each annotation): bers) that could be used by machines to compare them. D Φ→W (iY:1{},, ℜ . One could then look for the nearest image similar to a ) given text label or videos similar to a given image and In the simplest case, we assume that both mappings so on. Each type of object (image, video, music, text) are linear, that is: T could have a different transformation, or mapping func- Φ=I (xx) V tion, and these could be learned using appropriate train- and ing data. Φ=Wi(iW) , Figure 1 illustrates our idea for a simple case that involves two types of objects—images and text labels. We propose to learn two mappings (or transformations) jointly into a feature space (which we call embedding space) where the images and labels are both represented. Even though the mapping functions for images and text labels are different, they can be learned jointly to opti- mize the objective of our final task—annotating images with appropriate text labels. The first step consists in transforming images, which are often available in jpg, gif, and similar file formats, into more appropriate representations. Years of research in computer vision have yielded several alternatives. Most recent approaches start by detecting so-called interest points, which are positions in the image consid- ered important to analyze (e.g., Dalal and Triggs, 2005; Lowe, 1999). Features are extracted from these interest points using well-known methods, such as scale-invariant feature transforms (SIFT) or histograms of oriented gra- dients (HOG). Finally, the features are aggregated into a single fixed-length vector representing the image. Figure 1 Joint embedding space. The 22 BRIDGE where V is a matrix of size d × D mapping images into decompose the ranking task into several smaller tasks: the embedding space, W is a matrix such that line i, Wi loss =−1 fx+−+ fx is a D-dimensional vector that represents the label in ∑∑∑ yy(()) x y+ y− + position i in the dictionary, and T means “transpose” and applies to the matrix that precedes it. For example, where the notation |a|+ is a if a>0, 0 otherwise, and xT means that the lines and columns in matrix x are for each training image x, we want the score of each transposed. appropriate label y+ for that image to be higher than The goal for a given image is to rank all possible the score of any inappropriate label y – for that image by annotations so that the highest ranked annotations best a margin of at least one. Otherwise we will incur the describe the semantic content of the image. Consider corresponding loss. the following model: This loss can be minimized very efficiently on very T large datasets using stochastic gradient descent. In this fx((=Φ xiΦ=( xVT W T iI))W ) i case, one randomly selects a triplet (x, y+, y –) from the where the possible annotations indexed by i are ranked training set and computes its loss. If the loss is 0, then according to the magnitude of fi(x), largest first. nothing changes. Otherwise, one can compute the gra- dient of the loss with respect to the parameters (in this Image Labeling as a Learning-to-Rank Task case, V and W) and use it to modify the parameters in Labeling an image can be viewed as a ranking task. the direction opposite the gradient with some fixed but Given an image, one must order labels so that the ones small learning rate, which amounts to a slight improve- at the top of the list best correspond to the image and ment in the model with respect to that triplet. By doing the ones at the end of the list are unrelated to it. Vari- so repetitively, this procedure can be shown to converge ous learning-to-rank methods have been proposed in to the nearest local optimum of the global loss. the machine-learning literature over the years, some However, note that most of the chosen triplets will of which can scale to large datasets and some, such as yield a 0 loss (e.g., it very quickly becomes easy to say SVM-Rank, that cannot (e.g., Joachims, 2003). that a given image is more like a dolphin than a truck), Here is the simplest scalable approach. One can hence not improving the model, which can lead to very

TABLE 1 Nearest Labels in the Embedding Space Learned on the Web-data

Target Label Nearest Labels

barack obama barak obama, obama, barack, barrack obama, bow wow, george bush

david beckham beckham, david beckam, alessandro del piero, del piero, david becham

santa santa claus, papa noel, pere noel, santa clause, joyeux noel, tomte

dolphin delphin, dauphin, whale, delfin, delfini, baleine, blue whale, walvis

cows cattle, shire, dairy cows, kuh, horse, cow, shire horse, kone, holstein

rose rosen, hibiscus, rose flower, rosa, roze, pink rose, red rose, a rose

pine tree abies alba, abies, araucaria, pine, neem tree, oak tree, pinus sylvestris

mount fuji mt fuji, fujisan, fujiyama, mountain, zugspitze, fuji mountain

eiffel tower eiffel, tour eiffel, la tour eiffel, big ben, paris, blue mosque, eifel tower

ipod i pod, ipod nano, apple ipod, ipod apple, new ipod, ipod shuffle

f18 f 18, eurofighter, f14, fighter jet, tomcat, mig21, f 16

SPRING 2012 23

slow progress. A faster alternative would be a loss that images and more than 100,000 labels. The labels cor- concentrates on the top of the ranking instead of con- respond to queries people have tried on Google Image sidering all triplets (x, y+, y –) uniformly. Search, and images attributed to these labels were the In Weston et al., 2010, we proposed the weighted ones they often clicked on. This made for a very noisy approximate rank loss (WARP), which can weigh each dataset in which queries were written in several lan- triplet according to the estimated rank of appropriate guages with many spelling mistakes and many apparently labels and still yield an efficient implementation. In similar queries (although there are often subtle differ- order to estimate the rank of appropriate labels effi- ences between apparently similar queries, like dolphin, ciently, we used the following trick: for a given image which usually means the sea animal, and dolphins, which x and its correct label y+, we randomly sampled several often also means the football team in the United States). incorrect labels, from y – up, until we found one that An interesting side effect of training such a model yielded a non-negative loss. is that looking at the nearest labels to a given label in The number of samples that had to be drawn from the embedding space provides a natural way of orga- the set of possible labels to find one that yielded a non- nizing labels. Examples in Table 1, which show the negative loss is a good indicator of how well-ranked the nearest labels to a given label, include several mis- correct label is. If it is very well ranked, one will have spellings, translations, and semantically similar labels. to sample many incorrect labels before finding one that Table 2 provides examples of images from our test set is ranked higher. If it is poorly ranked, one will quickly of 3,000,000 along with the nearest 10 labels in the find an incorrect label that is ranked higher. embedding space. In both tables, one can see that the In fact, a good estimate of the rank of the correct model is not perfect (e.g., the image of the dolphin was label is simply the number of incorrect labels in the dic- labeled “orca” in position 2, before “dolphin” in posi- tionary divided by the number of samples needed in the tion 3). However, given the difficulty of the task, the proposed approach. One can then use this estimate to model works quite well. weigh the gradient update accordingly. The resulting model can be trained much faster and will perform at a Learning Label Trees much higher level of precision at the top of the ranking. Real-time applications are not effective for label- ing images when the number of labels is large (in our Large-Scale Learning testing dataset we had about 100,000 labels). A naive We trained an embedding model with the WARP loss approach would be to compute a score for each poten- on a very large dataset of more than 10,000,000 training tial label, then sort them, and finally show the top few.

TABLE 2 Examples of the Top 10 Labels Obtained for Some Test Images

Target Image Nearest Labels Target Image Nearest Labels

delfini, orca, dolphin, mar, barrack obama, barack obama, delfin, dauphin, whale, cancun, barack hussein obama, barack killer whale, sea world obama, james marsden, jay z, obama, nelly, falco, barack

eiffel tower, statue, eiffel, mole ipod, ipod nano, nokia, i pod, antoneliana, la tour eiffel, nintendo ds, nintendo, lg, pc, londra, cctv tower, big ben, nokia 7610, vino calatrava, tokyo tower

The 24 BRIDGE

In our model, we apply the following two steps: Obama Fish (1) learning a label tree, and (2) learning predictors for each node of the tree. These steps are described in the Sting Car following subsections. Dolphin Toyota Learning the Label-Tree Structure Whale Phone To learn a label tree such as the one in Figure 2, we proceed as follows. Given a set of labels in a node, we look for a partition of that set into subsets. Within the subset, labels are difficult to distinguish with clas- Obama Dolphin Car sifiers trained on their corresponding images. How- Sting Whale Toyota ever, it is easier to distinguish images belonging to labels of one subset from images belonging to labels of Fish Phone another subset. Obama Sting This basic rule enables us to postpone difficult deci- Dolphin sions (e.g., is this an ipod nano or an ipod touch) and Whale Phone make easy decisions first (e.g., is this about either [dol- phin, whale, shark] or about [ipod touch, ipod nano, Fish Car Toyota iphone]). We can do this by computing the confusion matrix between all labels to count the number of times our classifiers confuse class i with class j on a test set of Car Toyota images not used to train the classifiers. We then use the resulting matrix to apply spectral Figure 2 Example of a label tree. clustering (Ng et al., 2002), a technique that can cluster objects (in this case labels) using a given similarity (or However, as the number of potential labels increases, distance) matrix between them, instead of using, say, this procedure becomes infeasible in real time. Ideally, the Euclidean distance between two labels. This proce- the labeling time should be, at most, proportional to the dure can then be used recursively to create a complete log of the number of labels. label tree. Table 3 shows an example of labels that were Existing tree-based approaches (Beygelzimer et al., clustered together using this technique. 2009a,b, Hsu et al., 2009) are typically less accurate than naive linear-time approaches. We have proposed Learning a Label-Embedding Tree a novel approach for learning a label tree, in which each Once labels are organized into a tree, one can retrain node divides the set of potential labels into multiple a new embedding model (using the algorithm described subsets and a local decision is made to choose which above) in which each image can be labeled either with subset of labels the current image is most probably asso- its original set of labels or with reference to the nodes ciated with (Bengio, 2010). Figure 2 shows an example of the tree that contain them. Thus, the embedding of a label tree. If the subsets are balanced (i.e., approxi- space contains representations of all labels, as well as all mately the same size), this procedure decreases the nodes of the tree. number of labels to consider at a logarithmic rate until At each training iteration, one selects an image at a prediction is reached. random and then selects one of its correct labels at

TABLE 3 Examples of Obtained Clusters of Labels

great white sharks, imagenes de apple iphone 3gs, apple ipod, apple chevy colorado, custom trucks, dodge delfines, liopleurodon meduse, tablet, bumper, iphone 4, htc diamond, ram, f 250, ford excursion, ford f 150, mermaid tail, monstre du loch ness, htc hd, htc magic, htc touch pro 2, mini truck, nissan frontier, offroad, monstruo del lago ness, oarfish, oceans, iphone 2g, iphone 3, iphone 5g, pickup, toyota tundra sea otter, shark attacks, sperm whale, iphone app, iphone apple, iphone tauchen, whales apps, iphone nano

SPRING 2012 25

random, including the nodes of the tree where the References labels appear. Whenever an internal node is selected Bengio, S., J. Weston, and D. Grangier. 2010. Label embed- as a positive label for a given image, we select the com- ding trees for large multi-class tasks. In Advances in Neu- peting negative label as one of the sibling nodes in the ral Information Processing Systems, NIPS, 2010. label tree. This corresponds to how the tree would be Beygelzimer, A., J. Langfordm, and P. Ravikumar. 2009a. used at test time. Error-correcting tournaments. Pp. 247–262 in Interna- This repetitive procedure results in a structure of tional Conference on Algorithmic Learning Theory, ALT. labels based on both semantic and visual similarities. Beygelzimer, A., J. Langford, Y. Lifshits, G. Sorkin, and A. In addition, using this method, the performance of a Strehl. 2009b. Conditional Probability Tree Estimation label-embedding tree, on average, performs both faster Analysis and Algorithm. In Conference in Uncertainty in and slightly better at test time (Bengio et al., 2010). Artificial Intelligence, UAI, 2009. Dalal, N., and B. Triggs. 2005. Histograms of Oriented Conclusion Gradients for Human Detection. In IEEE Conference on The problem of image annotation for very large num- Computer Vision and Pattern Recognition, CVPR, 2005. bers of possible labels is very challenging. In recent Deng, J., W. Dong, R. Socher, L.-J. Li, K. Li, and L. Fei-Fei. years, significant progress has been achieved by propos- 2009. ImageNet: A Large-Scale Hierarchical Image Data- ing progressively better feature-extraction techniques base. In IEEE Conference on Computer Vision and Pattern from images to solve problems like rotation and transla- Recognition. tion invariances. In this article, we focus on the mod- Hsu, D., S. Kakade, J. Langford, and T. Zhang. 2009. Multi- eling aspect of the problem, particularly for very large, Label Prediction via Compressed Sensing. In Advances in varied datasets of images and labels. Neural Information Processing Systems, NIPS, 2009. The proposed approach (learning a joint embedding Joachims, T. 2003. Optimizing Search Engines using Click- space) achieves a performance level that is at least as through Data, Proceedings of the ACM Conference on high as performance levels on a similar task at that Knowledge Discovery and Data Mining, 2003. scale. In addition, the approach described above yields Lowe, D. 1999. Object recognition from local scale-invariant a novel visual semantic space in which images and text features. Pp. 1150–1157 in Proceedings of the Internation- labels coexist and their relative distances in that space al Conference on Computer Vision, Vol. 2. reflect both semantic and visual differences. Ng, A.Y., and M.I. Jordan, and Y. Weiss. 2002. On Spectral We also propose a method, an additional tree- Clustering: Analysis and an Algorithm. In Advances in based structure over the labels that learns from data, Neural Information Processing Systems, NIPS, 2002. so that a given image can be annotated quickly in Torralba, A., R. Fergus, and W.T. Freeman. 2008. 80 million a time frame proportional to the log of the number tiny images: a large dataset for non-parametric object and of potential labels. With its improved performance scene recognition. IEEE Transactions on Pattern Analysis and quick reactions, this approach may prove to be and Machine Intelligence 30(11): 1958–1970. very efficient. Weston, J., S. Bengio, and N. Usunier. 2010. Large Scale Image Annotation: Learning to Rank with Joint Word- Image Embeddings. In Proceedings of the European Con- ference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML-PKDD, 2010. Improving the construction and performance of buildings could lead to huge reductions in carbon emissions.

Challenges and Opportunities for Low-Carbon Buildings

John Ochsendorf

Buildings account for a higher proportion of greenhouse gas emissions than any other economic sector, and emissions attributable to building con- struction and operation have been increasing in recent decades. These emis- sions can primarily be traced back to heating, cooling, and lighting systems, although emissions from embodied materials are also significant. Major new initiatives are under way to reduce the amount of energy con- John Ochsendorf is associate sumed by buildings, but numerous technical, economic, and policy barriers will professor, Building Technology have to be overcome. This article provides summaries of some key challenges Program, Department of Civil to the design and implementation of low-carbon buildings in the United States and describes the resulting opportunities for the engineering professions. and Environmental Engineering, Department of Architecture, Introduction Massachusetts Institute of Although buildings have major environmental impacts in terms of water use, consumption of raw materials, and the depletion of other natural Technology. resources, the focus of this article is on reducing greenhouse gas emissions for mitigating climate change. Carbon dioxide equivalence (CO2e) provides a simple metric for determining the environmental performance of buildings, including both embodied emissions and operating emissions. As the largest source of carbon emissions in the United States, build- ings also represent a significant target of opportunity for carbon reductions (Figure 1). In a landmark paper, Pacala and Socolow (2004) identified SPRING 2012 27

improvements in building efficiency as one of seven “stabilization wedges” that could help offset global increases in carbon emissions over the next 50 years. The Intergovernmental Panel on Climate Change has identified buildings as the sector with the greatest potential for carbon reductions, particularly because reductions that result from improved building perfor- mance also yield substantial economic benefits (IPCC, 2007). The World Business Council for Sustainable Development has concluded that the energy use of buildings worldwide could be reduced by 60 percent by 2050 using existing technologies (WBCSD, 2009). Figure 1 U.S. energy consumption by sector. Source: Energy Information Agency. McKinsey Consulting has identified the building sector as the most cost-effective target for carbon abatement. systems; and (3) using off-site renewables, such as off- According to McKinsey’s analyses, carbon reductions shore wind energy (up to a maximum of 20 percent). for most buildings could be achieved at a negative cost Although the 2030 Challenge is voluntary at present, (McKinsey, 2007). many city governments have committed to pursuing its Current Initiatives goals, and a number of engineering and architectural firms are already tracking energy consumption and car- Several current initiatives in the United States bon emissions for their new buildings in relation to the have established target levels for improved carbon 2030 Challenge goals. performance of buildings. The best known initiative is the LEED (leadership in energy and environmental National and State Initiatives design) rating system developed by the United States Some countries are developing binding legal require- Green Building Council. The exponential growth of ments for dramatic carbon reductions in new buildings the LEED system has dramatically increased the market in the coming decades. For example, in the United share of green construction over the last decade. How- Kingdom (UK), the Climate Change Act of 2008 man- ever, although the LEED system promotes a variety of dates 80 percent carbon reductions on a 1990 baseline environmental improvement strategies for many dif- by the year 2050 (Crown, 2008). To help achieve this ferent types of buildings, it does not explicitly address goal, the target for new residential construction in the carbon emissions. UK will be zero carbon emissions by 2016. The most aggressive legal target for carbon reduc- The 2030 Challenge tions in the United States is California’s Assem- The 2030 Challenge developed by Architecture2030 bly Bill 32 (AB32), which mandates reductions to establishes rigorous targets for reducing carbon emis- 1990 emissions levels by 2020 (Assembly Bill 32, sions from new buildings, and in the next two decades, 2012). Improved building efficiency is a major compo- these standards will become increasingly stringent nent of California’s road map toward lower greenhouse (Architecture2030, 2012). The current goal is to design gas emissions. buildings that use 60 percent less annual energy than average for that building type. For example, to satisfy Improving Building Codes the 2030 Challenge, a new hospital must emit 60 per- Other policies, such as more stringent building cent lower emissions than the current national average codes, can be important motivators for engineers for hospitals. The reduction targets will decrease by 10 and architects and can move markets toward lower percent every five years (e.g., a decrease in emissions of carbon buildings. An extensive study by the World 70 percent by 2015), until, by 2030, new buildings will Business Council for Sustainability has shown that be carbon neutral. more stringent building codes are the most effec- There are three primary approaches to meeting tive way to reduce carbon emissions from buildings these design goals: (1) improving design strategies; and to transform the market for a low-carbon future (2) improving the efficiency of technologies and (WBCSD, 2009). The 28 BRIDGE

Challenges for Low-Carbon Buildings in construction are spent on R&D in the United States, The Role of Engineers much less than is spent by industries in other economic sectors, such as the automobile and electronics indus- Engineers, who are crucial to designing more sustain- tries (Gould and Lemer, 1994). Furthermore, only 0.2 able buildings, often become involved too late in the percent of federal research funding is spent on topics design process to make all of the necessary decisions. related to green buildings (USGBC, 2008). Many key decisions, such as building orientation, glazing The scarcity of research funding means that fewer ratio (i.e., area of glass/area of opaque wall), and the over- researchers are working on this vital topic, despite the all form of the building are made in the earliest design urgency of climate change and the favorable econom- stages. Once these critical decisions have been made, ics for carbon reductions through improved buildings. engineers can attempt to optimize a poor design, but it Increased research funding would not only attract more is difficult at that point to achieve a low-carbon design. students, but would also attract leading engineers to this important field. The engineering of Opportunities for Low-Carbon Buildings More efficient building design is one of the most sustainable buildings is not cost-effective opportunities for large-scale reductions in carbon dioxide emissions on a national and global being taught in most U.S. scale. Thus more emphasis on integrated building schools of engineering. design for the full life cycle of a building can lead to dramatic improvements in building performance. The key is to incorporate systems thinking at the conceptual design stage, taking into account climate-specific factors The challenge is to integrate engineering analysis and regional climate. in a way that provides rapid feedback to architects and The choice of materials for a building can not only the rest of the design team early in the process. For determine the embodied carbon of a building, but also this, engineers must be trained as designers, so they can has implications for the carbon emissions from building propose multiple solutions to open-ended problems. In operations, through thermal mass and improved day- short, the design of high-performance buildings requires lighting. For example, appropriate building materials integrated systems thinking beginning in the earliest can moderate diurnal temperature swings or help to dis- conceptual design stage. tribute daylight deeper into interior spaces. In addition, To ensure that engineers with the necessary skills decisions about building orientation, façades, heating are available, more of them must be trained in build- and cooling strategies, and glazing ratios, which must be ing science and sustainable design. However, most made early in the design process, are crucial factors in engineering schools do not directly address the design the final energy performance of a building. and operation of buildings, because sustainable building design involves aspects of mechanical engineering, civil New Design Tools engineering, and architecture. Most traditional architectural design software does not The few existing programs in architectural engineering have the capacity to analyze energy or environmental are turning out graduates, but in numbers far below those performance in the conceptual design phase. However,­ of traditional engineering disciplines. Thus, despite the in the last decade, several tools have been developed to dramatic economic and environmental impacts of build- assist architects and engineers by providing architects ings and the growing need for engineers in this field, the with rapid feedback on environmental performance. engineering of sustainable buildings is not being taught For example, the MIT Design Advisor enables the in most schools of engineering in the United States. rapid estimation of building energy use based on mass- Funding for Research ing, orientation, climate, glazing ratios, and other factors (Design Advisor, 2012). More recently, the DIVA plat- Finally, there is an acute lack of spending on research form has been developed to enable architects to run basic and development (R&D) for sustainable buildings. It performance simulations in the early design stage using has been estimated that only 0.25 percent of gross sales SPRING 2012 29

existing architectural design software (DIVA, 2012). addition, new graduate research fellowships for students Although such programs are being developed quickly, working in building science would help attract more there is still an urgent need for improvement, as well as researchers to the field. for engineers capable of systems-level building design. Case Studies Educating Architectural Engineers Many low-carbon buildings constructed in recent A multidisciplinary education focused on sustainable years can serve as models to inspire engineers, archi- design can attract a new generation of engineers to the tects, building owners, and policy makers. Noteworthy­ profession. In addition to expertise in heat transfer, ther- examples of a residential building development, a mal science, materials engineering, and other traditional school, an office building, and a cultural building are building sciences, sustainable buildings require expertise described below. in design thinking and creative problem solving. There is also a strong demand in the market for BedZED, Beddington, England, 2002 increased literacy in rigorous sustainability metrics and The Beddington Zero Energy Development, known for expertise in the life-cycle environmental and eco- as “BedZED,” is a pioneering low-carbon residential nomic performance of buildings. Integrated graduate development outside London, England, made up of 99 education in the built environment, such as the Solv- homes of various sizes (BedZED, 2012). Opened in ing Urbanization Challenges by Design, a program at 2002 and designed by a team of architects, engineers, Columbia University, combine engineering, architec- and developers, these low-rise buildings have passive ture, and planning to provide a broad-based educa- heating, cooling, and lighting strategies that dramati- tion for sustainable building designers and researchers cally reduce residential energy requirements (Figure 2). (IGERT, 2012). However, the BedZED concept goes beyond the design of buildings to include transportation, food pro- Funding for Research and Development duction, and other holistic urban-design strategies. For Increasing funding for R&D on sustainable buildings example, photovoltaic arrays and biofuel incinerators will require new partnerships among industry, academia, are designed to generate 100 percent of on-site electric- and government. Some current examples are (1) the ity demand. Greater Philadelphia Innovation Cluster for Energy Although the initial cost of construction was slightly Efficient Buildings supported by the U.S. Depart- higher than for conventional construction, occupants ment of Energy (DOE) and numerous academic and are realizing substantial annual energy savings, and the industry partners (GPIC 2012); (2) the Center for Architecture Science and Ecology supported by Rens- selaer Polytechnic Insti- tute and Skidmore Owings and Merrill (CASE 2012); and (3) the Advanced Building Systems Integra- tion Consortium at Car­ negie Mellon University (ABSIC, 2012). Increased federal funding could be provided by the National Science Founda- tion and the DOE, both of which are appropriate agen- cies for promoting research on sustainable buildings. In FIGURE 2 Beddington Zero Energy Development, Beddington, England, 2002. Source: WikiPedia “BedZED.” The 30 BRIDGE

in equally dramatic reduc- tions in life-cycle economic costs. Advanced building tech- nology in schools also has pedagogical benefits. For example, it can motivate students to learn about green building design and construction. Students and teachers also actively moni- tor the energy performance FIGURE 3 Richardsville Elementary School, Richardsville, Kentucky, 2010. Source: Sherman Carter Barnhart Architects. of the building, and green technologies are incorpo- resale value of homes has exceeded the local average rated into lesson plans. Richardsville is an example of housing market by 10 to 20 percent (BedZED Toolkit how an integrated design team can dramatically reduce II, 2003). Now 10 years old, BedZED continues to set the energy demands of a building to achieve a zero- an international standard for developers and designers energy school. to learn from and emulate. Research Support Facility, National Renewable Energy Richardsville Elementary School, Richardsville, Laboratory, Golden, Colorado, 2010 Kentucky, 2010 The Research Support Facility of the National Richardsville Elementary School is an 82,000 square Renewable Energy Laboratory is a 222,000 square foot, foot net-zero-energy educational building for 600 stu- net-zero-energy office building located at an altitude of dents (Sherman, 2012). Like other low-carbon build- 5,300 feet in the Rocky Mountains of Colorado (NREL, ings, Richardsville’s initial design was significantly more 2012). Home to 800 employees, this federal office efficient than the design of typical school buildings building is both a showcase for low-energy building (Figure 3). For example, annual energy demands are technology and a living laboratory for ongoing research only 5.3 kilowatt hours per square foot per year (kWh/ and for monitoring the performance of green technolo- ft2/yr), about 75 percent lower than the national stan- gies (Figure 4). dard for elementary schools. In addition, a 348 kW Equipped with day-lighting and passive heating and photovoltaic array generates on-site renewable energy cooling strategies, the building requires only 10.3 kWh/ to meet this demand. ft2/yr, which is 50 percent less than the national stan- At a cost of approximately $190 per square foot, the dard for office buildings. The energy demand is met school’s construction costs are competitive with those by an on-site 1.6 megawatt photovoltaic array, which of other schools around the country. Insulated concrete form (ICF) construction provides improved insula- tion and reduces air infil- tration through the exterior envelope of the building, and geothermal technology generates heating, cooling, and hot water. Overall, Richardsville has dramati- cally lower annual operat- ing costs than traditional school buildings, resulting FIGURE 4 Research Support Facility, National Renewable Energy Laboratory, Golden, Colorado, 2010. Source: NREL/ Dennis Schroeder. SPRING 2012 31

also helps power an on-site energy-intensive data cen- ter. Even at an altitude of one mile above sea level, the building does not require conventional heating. Instead, it is warmed by the data center and by trans- pired solar collectors, which were originally developed at NREL.

Interpretive Centre, Mapungubwe National Park, South Africa, 2009 The Mapungubwe National Park Interpretive Centre is both a museum and a visitor’s center at a World Heri- tage site in a remote area of northeast South Africa (Fig- ure 5). The client for the project, South Africa National FIGURE 5 Interpretive Centre, Mapungubwe National Park, South Africa, 2009. Parks, required a building with minimal environmental Source: James Bellamy. impact to be built, as much as possible, with local mate- rials and labor. In this harsh climate, where daytime Conclusions summer temperatures can reach 110°F, the design team Buildings are widely recognized as the most signifi- was determined to use passive strategies to maintain a cant opportunity for cost-effective carbon reductions comfortable interior temperature. in the United States and around the world. To take Using on-site soil and stones, the advanced structural advantage of these opportunities, engineers must con- design of the building consists of massive masonry walls tinue to play a vital role in transforming the built envi- and load-bearing vaults that increase thermal mass. The ronment for a low-carbon future. soil-cement vaults were built with minimal formwork Numerous areas for improvement remain. An inten- and no interior steel reinforcements, thus reducing the sive emphasis on conceptual design, life-cycle thinking, amount of embodied carbon in the building materials and innovative research partnerships will be necessary by approximately 80 percent compared to conventional to advance the field, reduce carbon emissions, and train steel and concrete construction (Ramage et al., 2010). a new generation of engineers. We now know from Because of the remote location, on-site energy experience that dramatic improvements in the design demands were kept to a minimum, and daylighting is and operation of buildings are possible. Our hope is the primary lighting source. For its radical approach to that well trained engineers will provide leadership for design and construction with local materials, the build- the United States and the carbon-constrained world of ing was honored as the 2009 World Building of the Year the future. at the World Architecture Festival. References Lessons Learned ABSIC (Advanced Building Systems Integration Consor- In these and most other low-carbon buildings, the tium). 2012. Advanced Building Systems Integration crucial design strategies were developed by collab- Consortium, Carnegie Mellon University. Available orative teams of architects and engineers early in the online at http://www.cmu.edu/architecture/research/cbpd/ design process. In addition, owners or clients often absic-cbpd.html. played a central role in establishing a vision for each Assembly Bill 32. 2012. California Global Warming Solu- project, essentially setting a challenge for the design tions Act. Available online at www.arb.ca.gov/cc/ab32/ team to meet. By setting clear targets for energy con- ab32.htm. sumption, these designs were achieved at costs similar Architecture2030. 2012. The 2030 Challenge. Available to those of conventional construction. However, as a online at http://www.architecture2030.org/2030_challenge/ result of reduced operating energy requirements, they the_2030_challenge. have significant life-cycle economic benefits. These BedZED. 2012. Beddington Zero Energy Development, and other projects clearly show that the carbon inten- Bioregional Development Group, Wallington, United sity of new buildings can be dramatically reduced using Kingdom. Available online at http://www.bioregional.com/ existing technologies. what-we-do/our-work/bedzed/. The 32 BRIDGE

BedZED Toolkit II. 2003. BedZED Toolkit II: A practical McKinsey. 2007. Reducing U.S. Greenhouse Gas Emis- guide to producing affordable carbon neutral develop- sions: How Much at What Cost? U.S. Greenhouse ments. London: Bioregional Development Group. Avail- Gas Abatement Mapping Initiative Executive Report. able online at http://www.bioregional.com/files/publications/ McKinsey and Company. http://www.mckinsey.com/en/ BedZED_toolkit_part_2.pdf. Client_Service/Sustainability/Latest_thinking/~/media/ CASE. 2012. Center for Architecture Science and Ecol- McKinsey/dotcom/client_service/Sustainability/PDFs/ ogy, Rensselaer Polytechnic Institute. Available online at Reducing%20US%20Greenhouse%20Gas%20Emissions/ www.case.rpi.edu. US_ghg_final_report.ashx. Crown. 2008. Climate Change Act 2008. Chapter 27. NREL. 2012. Sustainable NREL: Research Support Facility. United Kingdom: The Stationary Office Limited. Avail- Available online at http://www.nrel.gov/sustainable_nrel/ able online at http://www.opsi.gov.uk/acts/acts2008/pdf/ rsf.html. ukpga_20080027_en.pdf. Pacala, S., and R. Socolow. 2004. Stabilization wedges: Design Advisor. 2012. The MIT Design Advisor. Building solving the climate problem for the next 50 years with Technology Program, Massachusetts Institute of Technol- current technologies. Science 305(5686): 968–972. ogy. Available online at http://designadvisor.mit.edu. [DOI:10.1126/science.1100103] DIVA. 2012. Welcome to DIVA-for-Rhino. Available Ramage, M., J. Ochsendorf, and P. Rich. 2010. Sustainable online at www.diva-for-rhino.com. shells: new African vaults built with soil-cement tiles. Gould, J.P., and A.C. Lemer, eds. 1994. Toward Infrastruc- Journal of the International Association of Shell and Spa- ture Improvement: An Agenda for Research, Washington, tial Structures (IASS) 51(4): 255–261. D.C.: National Academies Press. Sherman. 2012. Going Green. Website for Sherman Carter GPIC (Greater Philadelphia Innovation Cluster). 2012. Barnhart Architects, Sustainable Design Initiatives. http:// Greater Philadelphia Innovation Cluster, U.S. Depart- www.scbarchitects.com/going-green. ment of Energy Innovation Hub. Available online at http:// USGBC (United States Green Building Council). 2008. gpichub.org/. A National Green Building Research Agenda. USGBC IGERT (Integrative Graduate Education and Research Research Committee, November 2007, Revised February Traineeship). 2012. Solving Urbanization Challenges by 2008. Available online at http://www.usgbc.org/ShowFile. Design. Columbia University IGERT Program. Available aspx?DocumentID=3402. online at http://www.civil.columbia.edu/igert/index.html. WBCSD (World Business Council for Sustainable Develop- IPCC. 2007. Climate Change 2007: Synthesis Report (AR4). ment). 2009. Transforming the Market: Energy Efficiency IPCC Fourth Assessment Report. New York: Cambridge in Buildings. Available online at http://www.wbcsd.org/ University Press. Available online at http://www.ipcc.ch/ Pages/EDocument/EDocumentDetails.aspx?ID=11006. pdf/assessment-report/ar4/syr/ar4_syr.pdf. The green building community is using information technology to drive fundamental changes in the real estate market.

Using Information Technology to Transform the Green Building Market

Christopher Pyke

The built environment shapes our lives both as individuals and as a society. It is not only a tangible library of our history, a reflection of our technol- ogy, and a window into our values and priorities, but also a platform for the future. The built environment greatly affects our health and well-being and is a major factor in our collective impact on the environment. Today’s buildings are marvels of security, comfort, and accessibility, but Christopher Pyke is vice president we are becoming increasingly aware of their limitations in terms of energy of research for the U.S. Green and water consumption, greenhouse gas (GHG) emissions, and storm water Building Council. runoff, as well as their contributions to diseases (e.g., asthma and obesity), exposure to toxic materials, and physical impairments. The impacts are felt both locally and over large geographic areas that are dependent on global supply chains. In fact, buildings are associated with 40 percent of global pri- mary energy consumption—50 percent if we include the energy consumed in the production and distribution of steel, concrete, aluminum, and glass (IPCC AR4 WG3, 2007). The negative impacts of buildings have historically been considered economic externalities, and, therefore, were unregulated and isolated from competitive forces in the real estate market. Nevertheless, their pervasive, often invisible, and unaccounted-for impacts have far-reaching implications. Ultimately, reducing the drivers of climate change, enhancing energy secu- rity, promoting public health, improving quality of life, and improving the The 34 BRIDGE health and well-being of current and future communi- As a result of efforts by thousands of organizations, ties will hinge on our ability to create (and re-create) the movement soon made a real impact on the building built environments. industry. Today, hundreds of thousands of profession- In the last 20 years, efforts by the green building com- als have received green building training and personal munity in the United States to respond to these chal- accreditation, and thousands of individual projects have lenges have been led by interdisciplinary, practitioner-led been evaluated and certified against third-party stan- coalitions seeking to advance the design, construction, dards (Watson, 2011). and operation of built environments to promote human The next generation of green building will return to health, well-being, and the protection and restoration its beginnings, propelled by an emerging foundation of of the environment. The movement began with an data and analytics that provide empirical support for the image and a simple idea, a classic market-adoption curve original vision. We can now cite data to describe the (a conceptual illustration of the distribution of prac- prevalence of performance and the distribution of actions tice or performance in an industry, ranging from a among market participants. We can create the “curves” few scofflaws on the low end through the average- we have been talking about and promoting, powered by performing majority, to a small group of innovators a state-of-the-art information ecosystem rooted in proj- on the high end [Roger, 1962]). The idea was to use ect performance and evidence-based practice. strategic market interventions to permanently shift dis- tributions of practice toward higher performance and Measuring Green Performance and Achievement greater achievement. We now have tools and processes for designing Although the concept attracted some attention, and assessing high-performance green buildings and the image lacked substance, mostly because very little communities. The concept of a green building “proj- information was available to describe distributions of ect” has evolved to encompass commercial interiors, practice and performance over relevant populations homes, commercial buildings, and even entire neigh- for real-world buildings. Empirical evidence for under- borhoods—ranging from less than 1,000 square feet to standing or predicting the real-world impacts of poten- more than 50 million square feet. With thousands of tial market interventions and green building practices projects completed and thousands more in the pipeline, was also scarce. We simply did not have the data or green building projects are becoming the currency of tools we needed to provide an empirical foundation to data-driven market transformation. support the founding vision. Recognition for projects is typically based on the implementation of specific design and operational strat- egies (e.g., integrative design, energy efficiency, water Green building projects conservation), the achievement of milestones (e.g., facilities management policies), and, increasingly, the are becoming the currency performance of whole systems, ranging from interior spaces to neighborhoods (e.g., whole-building energy of data-driven market performance).1 transformation. Project-rating systems, such as the U.S. Green Build- ing Council’s Leadership in Energy and Environmental Design (LEED™)2 and a number of analogous systems Fortunately, the lack of experience and data did not around the world define required levels of achievement delay action. Leaders of the movement first defined and performance to meet a wide range of criteria (Cole, “green” and set an initial direction by defining “good,” “better,” and “best” practices and preferred levels of 1 The term “performance” refers to a measurement, typically a quantita- performance. In this way, they provided a credible tive metric, such as energy efficiency, renewable energy generation, directional signal to the industry—not perfect and not water consumption, or occupant satisfaction. The term “achievement” refers to binary or qualitative activities, such as policies, procedures, always rooted in data, but useful as a basis for action. or discrete choices (e.g., green cleaning, commissioning, or the use of The next steps were to develop a workforce and assess third-party certified building products). The two terms are often used together as “performance and achievement” to reflect the typical range tools, policies, and programs for recognizing and reward- of green building practices. ing high performance. 2 See www.usgbc.org/leed for more information. SPRING 2012 35

1999). Projects that meet or exceed requirements are However, comprehensive information on non- recognized by LEED certification and other designations. financial dimensions of properties, such as energy use, In effect, assessment systems and third-party certifi- water consumption, and occupant experience, or on the cation tools have responded to historic deficiencies in efficacy of green building practices, such as energy effi- information and market failures (e.g., the lack of infor- ciency technologies and water-conserving products is mation about energy efficiency or GHG emissions) by not readily accessible. The absence of this information providing credible information about practices and per- contributes to inefficient markets, inhibits innovation, formance levels. For example, LEED certification brings and sometimes contributes to market failures (Gilling- transparency to many aspects of green building, including ham et al., 2009). location and transportation, design and engineering pro- cesses, construction activities, site planning, energy con- sumption and supply, water use, indoor environmental quality, and innovation. Such transparency improves the We can now treat individual flow of information among market participants and cre- projects and collections ates opportunities for competitive differentiation. Fun- damentally, the LEED standard provides the long-sought of projects as testable distribution of market conditions and rewards projects that demonstrate leadership in prevailing practices. hypotheses of the efficacy Over the past decade, the day-to-day tools for ensur- of green building processes, ing transparency have been rudimentary—a simple scorecard and, at the end of the process, a glass plaque products, and services. displayed in a building lobby. Nevertheless, these sim- ple steps have had a remarkable impact on the industry. Using information technologies, we can now vastly For example, inefficient, polluting assets may be accelerate and scale up the impact of our fundamen- overvalued, while cleaner, more efficient buildings are tal goals and concepts. In addition, we can now treat overlooked. The most effective practices may be under- individual projects and collections of projects as testable valued, and less effective actions may be overrated. hypotheses of the efficacy of green building processes, The resulting misallocation of resources is a predictable products, and services. Emerging information technol- result of the absence of information, and this is now the ogy and data sources provide opportunities for reinforcing defining challenge and opportunity for the green build- connections between design intent and real-world ing movement. impact. The remainder of this article describes a prelimi- The most direct remedy is to create public and pri- nary vision for information-powered, data-driven market vate mechanisms to provide information on non- transformation. financial aspects of assets, such as the green dimensions of homes and commercial buildings, along with information Information and Efficiency about the contributions of specific practices. This can be Classical economics begins with the assumption that accomplished through public labeling programs and pri- market participants have equal and immediate access vate efforts to create information markets and analytical to relevant information (Fama, 1970; Malkiel, 2003), services. Some public programs have been initiated, such which is the basis for setting prices and valuing assets. as the European Union’s building-level Energy Perfor- Real estate markets, for example, have developed mance Certificates, green building certifications, and, in sophisticated tools that provide information on finan- a few metropolitan areas, municipal energy benchmarks cial aspects of individual buildings and portfolios, and (IEA, 2010). Complementary private efforts may also be commercial information services provide benchmarking in the offing, but only a few have emerged publicly (e.g., data related to capital costs, sales prices, tenancy, and Bloomberg New Energy Finance). numerous other factors. Markets for this information Taken together, however, these efforts barely scratch are mature and highly segmented, and market responses the surface. Ultimately, we must connect information are complex, but reasonably predictable (Gatzlaff and about performance with essential contextual infor- Tirtiroglu, 1995). mation in three essential areas: design intent (e.g., The 36 BRIDGE original plans), implementation actions and practices information technologies are creating opportunities (e.g., technologies, management strategies, etc.) and for scalable, information-based market interventions. normalization factors (e.g., occupancy schedules, occu- However, realizing these opportunities will require three pant density, etc.). key components in the design of emerging data streams and information technologies for green buildings: Accelerating Innovation • Define outcomes (i.e., performance dimensions sub- Real-world data on how operational performance ject to evaluation). “Green building” per se is not an relates to design intentions and specific strategies, outcome. It is a process and an evaluation framework. people, products, and services will be the foundation The next generation of assessments must strengthen and currency for the next generation of green buildings. connections between desired outcomes (intentions) However, data per se will not necessarily lead to change. and operational performance measures, which will To accelerate market transformation, information must lead to more specificity in design, engineering, opera- be interpreted and linked to mechanisms that create tions, and evaluations. market opportunities for high-performing projects and, by extension, competitive risks for low performers. • Understand and reward relatively high performers. At this point, the interests of the green building Data-driven evaluations of new performance dimen- movement diverge from those of agnostic market ana- sions can provide a basis for sorting and ranking proj- lytics. Rather than improving reporting on the status ects; drawing inferences about underlying practices, quo, our goal is to use information to drive perma- products, and services; and creating performance- nent, self-sustaining change in entire industries. Suc- based reward systems. cess will be measured by the rate and magnitude of • Inspire and assist relatively low performers. Infor- those changes. mation creates opportunities to identify and improve We intend to use data and information technology low-performing projects and strategies in comparison to accelerate the diffusion of innovation and the rate of to higher performing peers. market evolution, that is, the rate at which new prac- tices are taken up by market participants and, critically, Outcomes their impact on the performance of individual assets For the past decade, green building has been deter- and the built environment as a whole. Our goal is to mined by a single performance dimension—the number achieve a permanent, positive shift in prevalent prac- of points a project achieves on a rating system. This tices and performance. dimension is typically divided into categories, such Although the building sector as a whole has not as LEED Certified, Silver, Gold, and Platinum levels. yet embraced these concepts, new data sources and Sometimes, the act of certification itself or the level of certification becomes a goal in itself. Outcome 1 Outcome Outcome n We have been exploring ways to expand the tra- ditional focus by developing and implementing a multi­dimensional framework linking green building outcomes and practices. An initial framework was Strategy 1 (credit) released with LEED 2009 (USGBC, 2008), in which every green building “credit” (a.k.a., optional strategy) Strategy (credit) was quantitatively associated with 13 environmental “impact categories,” such as GHG emissions, resource

Strategy n (credit) depletion, and smog formation (Figure 1). On this basis, weights (points) were assigned to individual cred- its. In addition, credit achievement could be used to track specific outcomes—literally unpacking the infor- mation collected during the certification process. Figure 1 LEED is an outcome-oriented rating system. Points (relative weights) LEED 2012, which has seven core green building are assigned based on the association between credits (a.k.a., strategies) and outcomes supported by more than 30 metrics (e.g., outcomes (e.g., reductions in greenhouse gas emissions). energy efficiency, renewable energy production), will be SPRING 2012 37

designed from the bottom up to associate actions with outcomes. In addition, every action can be critically Distribution of performance across the market evaluated with respect to operational performance. The outcomes or performance dimensions can be as “simple” as intensity of energy use (e.g., annual energy use per square unit of floor space) or much more complicated, synthetic measures, such as the 29 weighted factors

included in the LEED 2009 GHG Index. Each metric Frequency provides a new dimension for ranking and sorting green How? building projects in terms of goals and outcomes. Increasing rewards

High Performers Low Outcome High Each performance dimension is populated with real Performance Performance projects based on third-party verified data collected dur- Figure 2 The Green Building Information Gateway (GBIG) provides data on the ing the certification process giving us an opportunity to distribution of performance across populations of green building projects. GBIG identify and reward high performers. The first test is information can be used to improve our understanding of “how” high-performing simple scoring based on performance (e.g., an Energy projects deliver above average results. Placing a project in context provides a basis Star score). However, the confluence of service-based, for recognition and competitive advantage. information technology architectures, highly scalable databases, and pervasive data collection is creating information systems to collect data on performance, unprecedented opportunities for connecting design practices, and technologies in near real time and pro- and engineering intent with execution and outcomes. vide dynamic, context-relevant benchmarking and Based on these data, we will be able to identify factors recommendations. These data will provide decision that contribute to different levels of performance and makers with timely information for market and green achievement (Figure 2). “comparables,” which are not currently available in the Fundamentally, we want to understand how projects real estate industry. achieve a given level of relative performance. This will require identifying and tracking relationships among people, organizations, practices, technologies, and many other factors. Each high-performing project can then provide a benchmark for lower performing proj- ects. Data on projects can be combined to evaluate the relative efficacy of green building practices. Today, we can use a demonstration system called the Green Building Information Gateway (GBIG) (www.gbig.org) to begin to identify and explore high- performing projects based on multiple outcomes. For example, Table 1 shows the performance and achieve- ment in six categories of an exemplary office building in Chicago, Illinois. The accompanying density plots compare the selected project (the dark triangle) with other projects certified on the same rating system, in this case LEED for Existing Building: Operations & Maintenance (more information is available from TABLE 1 A Basic “Nutrition Label” for an Exceptionally High-Achieving LEED- http://www.gbig.org/projects/10049661). Certified Project. Conceptually similar to food labeling, the “nutrition label” can be Our goal is to use this data and information tech- created by GBIG to summarize information for six LEED credit categories and place nology to shorten cycles between innovations, market achievement in the context of all others using the same rating system. Source: uptake, operational performance, and positive rec- Green Building Information Gateway (www.gbig.org) or http://www.gbig.org/ ognition. This will require creating highly scalable projects/10049661. The 38 BRIDGE

products, and services creates immediate opportunities Distribution of performance across the market for encouraging constructive competition and competi- tive differentiation. It also contributes to a new frame- work for testing and evaluating green building strategies using projects as natural experiments. However, realizing these opportunities will require a change in perspective. Today, the implementation of Frequency Why? strategies is often considered a mark of achievement—an

Increasing risks outcome in its own right. However, in the near future the value of green building strategies will be based on the reli- Low Outcome High ability of their contributions to operational performance Performance Performance and intended outcomes. In this emerging, performance- Figure 3 Information in GBIG can be used to understand “why” relatively low- oriented perspective, every project becomes a real-world performing projects deliver below average results. This understanding can inform experiment representing a unique combination of cir- critical evaluation and subsequent improvements. cumstances, intentions, strategies, and outcomes. Many sectors have embraced such frameworks, some- Low Performers times called evidence-based practices, adaptive man- The green building community has always been agement, or continuous improvement (e.g., the familiar comfortable recognizing high performers, but for every “Deming Cycle”). Typically, the goal is to treat prac- high performer there are a commensurate number of tice as iterative experimentation. Every action has an underperformers. Outside of Lake Woebegone, under­ intent, implementation strategy, and anticipated out- performers are statistically inevitable. comes. The objective is to incrementally strengthen Nevertheless, we have generally been less aggressive connections between intentions and outcomes. in searching out underachievers and trying to under- For example, some physicians practice evidence- stand and assist them. To ensure continued progress, based medicine by using data on health outcomes to we recognize that we must pursue an understanding of inform treatment decisions. Similarly, conservation these projects with energy equal to or greater than our biologists practice adaptive management using informa- pursuit of high performers. tion about population trends to refine restoration and Fortunately, we can adapt the same basic information technologies to identify underperforming projects (Fig- ure 3), determine the factors that affect their practices, and recommend specific strategies for improvement based on practices by comparable higher perform- ing projects. We may also be able to identify specific actions and circumstances that compromise outcomes. Our goal is to understand the challenges and, if neces- sary, create new or improved interventions to overcome barriers, such as technological limitations, lack of tech- nical understanding, or cost. GBIG can help identify and explore many outcome dimensions for relatively low-achieving projects. Table 2 shows selected metrics for a LEED for New Construction (version 2.2) project in Washington, D.C. (more infor- mation is available from http://www.gbig.org/projects/ 10100317). TABLE 2 A Sample GBIG “Nutrition Label” for a Relatively Low-Achieving LEED-Certified Project. This label places LEED credit category achievement in Testing and Evaluation the context of all other certified projects using the same rating system. Source: The ability to understand, stratify, rank, and interpret Green Building Information Gateway (www.gbig.org) or http://www.gbig.org/ performance and outcomes with respect to processes, projects/10100317. SPRING 2012 39

Figure 4 Proportion of LEED credits associated with different evaluation methods and rating systems. The x-axis categories reflect three LEED rating systems: LEED for New Construction, LEED for Operations and Maintenance, and LEED for Neighborhood Development. The y-axis indicates the fraction of prerequisites and credits intents that can be directly evaluated in operation based on physical measures (e.g., meters), occupant experience, or “not testable” (e.g., reference to a third-party standard or criteria). Source: Data from Pyke et al., 2010. protection practices. Evidence-based practice in these of a specified policy (Figure 4). The intentions of the domains is supported by closely coordinated monitor- other 50 per cent of LEED credits require evaluations of ing, evaluation, and surveillance programs. operations, most often using information from physical The architectural community is just beginning to rec- measures or human experience in or around projects. ognize similar opportunities for “evidence-based design” Physical measures include energy use, energy produc- (Mortice, 2009). Unfortunately, this industry gener- tion, water consumption, temperature, runoff, and other ally lacks corresponding systematic data collection and factors. Human experience reflects the health and well- systems, and interconnected data and analytics are pre- being of occupants, which can be measured through requisites for evidence-based practice. This community physiological responses or opinions and perceptions. will require an information infrastructure like the one Every time we link design or engineering intent and envisioned for GBIG. real-world actions with outcomes, we create new per- Green building will help lead this revolution. Every formance dimensions that can be used to stratify and credit in every LEED rating system is defined with compare projects. Combinations of intent, actions, respect to a specific intent and documentation. Today, and outcomes are enabled by search, compare, and more than 1,000 credits are in regular use for nearly benchmarking information technologies. The integra- 100,000 projects, representing real-world opportunities tion of these new streams of data and analytical tools for testing a wide range of hypotheses, some of which will provide a foundation for the next generation of are more testable than others (Pyke et al., 2010). performance-oriented, evidence-based practices. In a recent study, we found that 50 per cent of Clearly, we will have to place more emphasis on the LEED credit intents are currently evaluated based on management of information for the entire life cycle of consistency with a third-party authority or rule, such a project, design and engineering intentions, documen- as purchase of a certified material or implementation tation of actions and implementation strategies (e.g., The 40 BRIDGE service providers, products, etc.), and performance Gatzlaff, D.H., and D. Tirtiroglu. 1995. Real estate market monitoring and evaluation systems explicitly linked to efficiency: issues and evidence. Journal of Real Estate Lit- intended outcomes. All of these elements will be built eratue 3(2): 157–189. into increasingly integrated information ecosystems Gillingham, K., R. Newell, and K. Palmer. 2006. Energy that span the life cycle of built environments and the efficiency policies: a retrospective examination. Annual length and breadth of the supply chains that support Review of Environment and Resource Management 31: them. This emerging ecosystem will support a pervasive 161–192. network of data producers and information consumers Gillingham K., R.G. Newell, and K. Palmer. 2009. Energy seeking to create value and gain competitive advantage efficiency economics and policy. Resources for the Future through superior performance. Report No. RFF-DP-09-13. IEA (International Energy Agency). 2010. Energy Perfor- Conclusion mance Certification of Buildings. International Energy The success of green building over the past decade Administration Policy Pathway Report. Available at http:// attests to the efficacy of using relatively simple www.iea.org/papers/pathways/buildings_certification.pdf. information-based interventions to produce demon- Intergovernmental Panel on Climate Change (IPCC) strable market transformations. In the coming decade, Fourth Assessment Report (AR4) Working Group III we will need new tools and approaches to bring these (WG3) (2007), B. Metz, O.R. Davidson, P.R. Bosch, R. concepts to scale and to generate the pace of change Dave, and L.A. Meyer, eds. Climate Change 2007: Mitiga- necessary to achieve our mission of creating sustain- tion of Climate Change, Contribution of Working Group able, healthy, high-performance built environments. III to the Fourth Assessment Report of the Intergovern- I believe this change will be powered by a new gen- mental Panel on Climate Change. Cambridge University eration of data and information technologies specifi- Press, ISBN 978-0-521-88011-4 (pb: 978-0-521-70598-1). cally designed to promote market-based competition Malkiel, B.G. 2003. The efficient market hypothesis and its across multiple performance dimensions, to understand critics. Journal of Economic Perspectives 17(1): 59–82. and learn from high performers, and to recognize and Mortice, Z. 2009. Evidence-based design: the deeper meaning improve low performers. Every performance dimension to sustainability, building performance, and everything else. we track provides an opportunity for competitive dif- AIArchitect 16. June 26, 2009. Available online at http:// ferentiation. Every high-performing project we identify info.aia.org/aiarchitect/thisweek09/0626/0626b_evidence.cfm. and rank provides an opportunity for learning, recog- Nadel, S., J. Thorne, H. Sachs, B. Prindle, and R.N. Elliott. nition, and reward. Every low-performing project we 2003. Market transformation: substantial progress from a touch provides an opportunity for education, invest- decade of work. American Council for an Energy-Efficient ment, and improvement. Economy, Report No. A036. The critical technologies, such as customized search Pyke, C.R., S. McMahon, and T. Dietsche. 2010. Green engines, distributed sensors, social media, service-based Building & Human Experience: Testing Green Build- software architectures, and cloud solutions, are either ing Strategies with Volunteered Geographic Information. already in hand or are rapidly emerging. With these Research Program White Paper. June 10, 2010, 14 pages. tools, we will be able to engage orders-of-magnitude more Roger, E.M. 1962. Diffusion of Innovations, 1st edition. projects and, ultimately, move from episodic certification New York: The Free Press. to continuous performance and real-time monitoring. UNEP SBCI (United Nations Environment Programme, We can already see the contours of this new world, its Sustainable Buildings and Climate Initiative). 2010. sweeping implications for green building practices, and Common Carbon Metric for Measuring Energy Use and the built environments we may someday inhabit. Reporting Greenhouse Gas Emissions from Building Oper- ations. Available online at http://www.unep.org/sbci/pdfs/ References Common-Carbon-Metric-for_Pilot_Testing_220410.pdf. Cole, R.J. 1999. Building environmental assessment meth- USGBC (U.S. Green Building Council). 2008. LEED ods: clarifying intentions. Building Research and Informa- 2009 Credit Weighting. Washington, D.C.: USGBC. tion 27(4/5): 230–246. Available online at http://www.usgbc.org/DisplayPage. Fama, E. 1970. Efficient capital markets: a review of theory aspx?CMSPageID=1971. and empirical work. Journal of Finance 25(2): 383–417. Watson, R. 2011. Green Building Market and Impact Report. GreenBiz Group. 50 pages. Brain-computer interfaces create alternate communication channels for people with severe motor impairment.

The Evolution of Brain-Computer Interfaces

Eric C. Leuthardt

The ability to enable the brain to control an external device with thoughts alone is emerging as a real option for patients with motor dis- abilities. The goal of this area of study, known as neuroprosthetics, is to create devices, known as brain-computer interfaces (BCIs), that can acquire brain signals and translate them into machine commands that reflect the intentions of the user. In the past 20 years, neuroprosthetics has pro- Eric C. Leuthardt is a neurosurgeon gressed rapidly from fundamental neuroscientific discovery to initial trans- and assistant professor, Department lational applications. of Biomedical Engineering and Seminal discoveries beginning in the 1980s demonstrated that neurons in motor cortex, when taken as a population, can predict the direction and Department of Neurological speed of arm movements in monkeys (Georgopoulus et al., 1982, 1986; Surgery, Washington University Moran and Schwartz, 1999b). In subsequent decades, these findings were School of Medicine, St. Louis, translated into increasing levels of brain-derived control in monkeys and then to preliminary human clinical trials (Hochberg et al., 2006; Taylor et Missouri. al., 2002). In recent years, an emerging understanding of how cortex encodes motor and non-motor intentions, sensory perception, and the role of cortical plas- ticity in device control have led to new insights into brain function and BCI applications. These new discoveries have greatly increased the poten- tial of neuroprosthetics in terms of control capability and the variety of patient populations that can be helped. This article provides an overview The 42 BRIDGE of current BCI modalities and emerging research on the a cursor toward a target). Thus, the brain must change use of motor and non-motor areas for BCI applications its signals to improve performance. and assessments of their potential clinical impact. In addition, a BCI may also be able to adapt to the changing milieu of the user’s brain to further optimize Brain-Computer Interfaces: functioning. This dual adaptation requires a certain Definition and Essential Features level of training and a learning curve, for both the user A BCI is a device that can decode human intent and the computer. The better the computer and sub- from brain activity alone, thus creating an alternate ject are able to adapt, the shorter the training period to communication channel for people with severe motor achieve control. impairment. Explicitly, a BCI does not require the There are four essential elements to the practical “brain’s normal output pathways of peripheral nerves functioning of a BCI platform (Figure 1). All four ele- and muscles” to facilitate interaction with the environ- ments must work in concert to manifest the user’s inten- ment (Wolpaw et al., 2000, 2002). A real-world exam- tion (Wolpaw et al., 2002): ple would be a quadriplegic subject who can control a • signal acquisition, the BCI system’s recorded brain cursor on a screen with signals derived from individual signal or information input neurons recorded in primary motor cortex without the need for overt motor activity. It is important to empha- • signal processing, the conversion of raw information size that a true BCI creates a completely new output into a useful device command pathway for the brain. • device output, the overt command or control func- As a new output pathway, the user must have feed- tions administered by the BCI system back to improve how he or she alters electrophysiologi- cal signals. Similar to the development of a new motor • operating protocol, the manner in which the system skill (e.g., learning to play tennis), there must be con- is altered and turned on and off tinuous alteration of the subject’s neuronal output. This requires that the output be matched against Signal Acquisition feedback from the intended actions so that the output Signal acquisition is a real-time measurement of (swinging the tennis racket or altering a brain signal) the electrophysiological state of the brain. Brain activ- can be tuned to optimize performance to reach the ity is usually measured by voltage changes recorded via intended goal (hitting the ball over the net or moving

FIGURE 1 Essential features and components of a BCI. There are four essential elements to the practical functioning of a BCI platform: (1) signal acquisition, the BCI system’s recorded brain signal or information input; (2) signal processing, the conversion of raw information into a useful device command; (3) device output, the overt command or control functions administered by the BCI system; and (4) operating protocol, the manner in which the system is turned on and off and the way the user or a technical assistant adjusts parameters of the previous three steps to convert intentions to machine commands. All four elements are essential to manifesting the user’s intention (Schalk et al., 2004a). Source: Leuthardt et al., 2009. Reprinted with permission. SPRING 2012 43

electrodes, either invasive (under the skin) or non- invasive (on the surface of the skin). Theoretically, other measurements, such as blood flow detected with magnetic resonance imaging (MRI), altered magnetic fields measured by magnetoencephalog- raphy (MEG), and opti- cal signals, might be used for signal acquisition. FIGURE 2 Signals for a BCI. The figure shows the three general categories of signals for BCI application and their anatomic However, none of these location relative to the brain and its respective covering layers. EEG = electroencephalography; ECoG = electrocorticography. non-electrical signals is Source: Leuthardt et al., 2009. Reprinted with permission. currently practical or fea- sible for clinical application. The impetus for all of these analyses is to determine The most common types of signals include electro­ the relationship between an electrophysiologic event encephalography (EEG), electrical brain activity recorded and a given cognitive or motor task. For example, after from the scalp (Elbert et al., 1980; Farwell and Don- recordings are made from an ECoG signal, the BCI sys- chin, 1988; Freeman et al., 2003; Pfurtscheller et al., tem must recognize that an alteration of the signal has 1993; Sutter, 1992; Vidal, 1977); electrocorticography occurred in the electrical rhythm (feature extraction) (ECoG), electrical brain activity recorded beneath and then associate the change with a specific cursor the skull (Leuthardt et al., 2004, 2005; Schalk et al., movement (translation). As mentioned above, signal 2004b); field potentials, electrodes that monitor brain processing must be dynamic to adjust to the changing activity from within the parenchyma (Andersen et al., internal signal environment of the user. 2004); and “single units,” microelectrodes that moni- tor the firing of individual neuron action potentials Device Output (Georg­opoulus et al., 1986; Kennedy and Bakay, 1998; Overt action (actual device output) is then taken by Laubach et al., 2000; Taylor et al., 2002). the BCI. As in the previous example, this can result in Figure 2 shows the relationship between various sig- moving a cursor on a screen, choosing letters for com- nal platforms in terms of anatomy and the population munication, controlling a robotic arm, driving a wheel- sampled. Once acquired, signals are digitized and sent chair, or controlling an intrinsic physiologic process, to the BCI system for further interrogation. such as moving a limb or controlling bowel and bladder sphincters (Leuthardt et al., 2006b). Signal Processing In the signal processing stage of BCI operation, Operating Protocol there are two essential functions: feature extraction An important consideration for practical applications and signal translation. The first extracts significant of BCI devices is the overall operating protocol. This identifiable information from the gross signal; the refers to the manner in which the user controls how the second converts that identifiable information into system functions. The how includes, for example, turn- device commands. ing the system on or off, controlling the kind of feed- Converting a raw signal into a meaningful one back and how fast it is provided, controlling the speed at requires an array of analyses, from assessment of fre- which the system implements commands, and switch- quency power spectra, event-related potentials, and ing between device outputs. cross-correlation coefficients for analysis of EEG/ECoG The way the system functions is critical for BCI signals to the directional cosine tuning of individual functioning in real-world applications. In most cur- neuron action potentials (Levine et al., 2000; Moran rent research protocols, the parameters are set by the and Schwartz, 1999b; Pfurtscheller et al., 2003). investigator. In other words, the researcher turns the The 44 BRIDGE system on and off, adjusts the speed of interaction, and velocity of hand movements. This may be an impor- defines very limited goals and tasks. These are all things tant limitation, because signals associated with specific the user will eventually have do by him(her)self in an movement parameters are typically used in BCI systems unstructured applied environment. based on action-potential firing rates. Another limitation of EEG recordings is that the Current BCI Platforms detected amplitudes are very small. This makes them To date, three general categories of BCI platforms susceptible to artifacts created by sources outside the have been put forward as possible candidates for clini- brain, such as electromyographic (EMG) signals pro- cal application: EEG-based systems; single unit systems; duced by muscle contractions. and ECoG-based systems, an intermediate modality. Despite these limitations, EEG-based BCIs have been The categories are primarily determined by the source shown to support higher performance than might be from which the controlling brain signal is derived—the expected, including accurate two-dimensional (McFar- scalp, intraparenchymal neurons, or the cortical sur- land et al., 2008b; Wolpaw and McFarland, 2004) and face. The current status of each of these platforms is even three-dimensional control of a computer cursor described below in terms of level of control, surgical (McFarland et al., 2008a). To date, the large majority considerations, and clinical population. of clinical applications of BCI technologies for people with severe motor disabilities have been demonstrated using EEG (Kübler et al., 2005; Nijboer et al., 2008; Vaughan et al., 2006). EEG is currently the only Ultimately, however, the intrinsic lack of signal modality that has been shown robustness may have important implications for chronic applications of BCI systems in real-world environments. to actually help people BCI systems based on EEG typically require substantial training (Birbaumer, 2006; Wolpaw and McFarland, with paralysis. 2004) to achieve accurate one- or two-dimensional device control (about 20 and 50 30-minute training sessions, respectively), although some studies have Electroencephalography-Based Systems reported shorter training requirements (Blankertz et al., EEG-based BCIs derive brain signals from electrical 2006). Nevertheless, noise sensitivity and prolonged activity recorded from the scalp (Birbaumer et al., 1999; training are fundamental limitations in the widespread Blankertz et al., 2006; Farwell and Donchin, 1988; clinical application of EEG-based BCIs. Kübler et al., 2005; McFarland et al., 1993, 2008b; In summary, EEG has been shown to support much Millan et al., 2004; Muller et al., 2008; Pfurtscheller higher performance than was previously assumed and et al., 1993, 2000; Sutter, 1992; Vaughan et al., 2006; is currently the only modality that has been shown to Wolpaw et al., 1991; Wolpaw and McFarland, 1994, actually help people with paralysis. However, because of 2004). EEG-based BCIs are the most common for stud- significant limitations, it is currently not clear to what ies in humans, probably because this recording method extent EEG-based BCI performance, in the laboratory is convenient, safe, and inexpensive. and in clinical settings, can be improved. EEG provides relatively poor spatial resolution, because a large brain area must be involved to generate Single Neuron-Based Systems detectable signals (Freeman et al., 2003; Srinivasan et From a purely engineering point of view, the opti- al., 1998). Despite this limitation, signals relevant to mal method of extracting electrical information from BCI research can still be obtained from EEG, includ- the brain would be to place a series of small recording ing modulations of mu (8–12 Hz) and beta (18–25 electrodes directly into the cortical layers (1.5–3 mm) Hz) rhythms produced by sensorimotor cortex. These to record signals from individual neurons. In essence, rhythms show non-specific changes (typically decreases this is what single-unit action-potential BCI systems in amplitude) related to movements and movement do. Such systems have been very successful for limited imagery, but they do not provide specific information time periods in both monkeys (Carmena et al., 2003; about the details of movements, such as the position or Serruya et al., 2002; Taylor et al., 2002; Velliste et al., SPRING 2012 45

2008) and humans (Hochberg et al., 2006; Kennedy limitations for large-scale clinical application, either and Bakay, 1998). because of prolonged user training and poor signal-to- To extract single-unit activity, small microelectrodes noise limitations with EEG or because of the inability with ~20 micron diameter tips are inserted into the brain of single-unit constructs to maintain a consistent signal parenchyma where relatively large (e.g., 300 microvolt) (Bjornsson et al., 2006; Szarowski et al., 2003; Wolpaw extracellular action potentials have been recorded from and McFarland, 2004). individual neurons 10–100 microns away. These sig- ECoG-based systems might be an ideal trade-off nals are usually band-passed filtered from 300–10,000 for practical implementation (Leuthardt et al., 2005). Hz and then passed through a spike discriminator to Compared to EEG, the ECoG signal is substantially measure time intervals between spikes. more robust. Its magnitude is typically five times larg- The firing rates of individual neurons are computed er, its spatial resolution is much greater (0.125 versus in 10 to 20 millisecond bins and “decoded” to provide 3.0 cm for EEG), and its frequency bandwidth is sig- high-fidelity control of either a computer cursor or robot nificantly higher (0–500 Hz versus 0–40 Hz for EEG) endpoint kinematics (Georgopoulus et al., 1986; Moran (Boulton et al., 1990; Freeman et al., 2003; Srinivasan and Schwartz, 1999a; Wang et al., 2007). Given its et al., 1998). The latter quality, access to higher fre- high spatial resolution (100 microns) and high temporal quency bandwidths, promises particularly useful infor- resolution (50–100 Hz), this modality arguably provides mation for BCI operation (Gaona et al., 2011). the highest level of control in BCI applications. Many studies have demonstrated that different fre- Unfortunately, there are two major problems with quency bands carry specific, anatomically distinct infor- single-unit BCIs. First, the electrodes must penetrate into mation about cortical processing. Lower frequency the parenchyma where they cause local neural and vascu- bands, known as mu (8–12 Hz) and beta (18–26 Hz), lar damage (Bjornsson et al., 2006). This can initiate a which are detectable with EEG, are thought to be pro- cascade of reactive cell responses, typically characterized duced by thalamocortical circuits. These bands show by activation and migration of microglia and astrocytes broad anatomic decreases in amplitude in association toward the implant site (Bjornsson et al., 2006). with actual or imagined movements (Huggins et al., Second, single-unit, action-potential microelectrodes 1999; Levine et al., 1999; Pfurtscheller et al., 2003; are very vulnerable to encapsulation. The continued Rohde et al., 2002). presence of electrodes promotes the formation of a sheath composed partly of the reactive astrocytes and microglia (Polikov et al., 2005; Szarowski et al., 2003). This reactive sheath can have numerous deleterious ECoG-based systems might effects, including neural cell death and tissue resistance be an ideal trade-off for that electrically isolates the device from the surround- ing neural tissue (Biran et al., 2005; Szarowski et al., practical implementation. 2003; Williams et al., 2007). Research into novel biomaterial coatings and/or local drug delivery systems that may reduce the foreign body The higher frequencies, also known as gamma activ- response to implanted electrodes is ongoing. So far, ity, are only appreciable with ECoG. Gamma activity, however, the results are far from clinical application which is thought to be produced by smaller cortical (Abidian and Martin, 2008; Seymour and Kipke, 2007; assemblies, shows close correlation with the action- Spataro et al., 2005). The development of a long-term potential firing of tuned cortical neurons in primary BCI system based on single-unit activity may be delayed motor cortex in monkey models (Heldman et al., 2006). until these issues have been resolved. In addition, high-frequency changes have been associ- ated with numerous aspects of speech and motor func- Electrocorticography-Based Systems tion in humans (Chao et al., 2010; Crone et al., 1998, Over the past five years, enthusiasm has been mount- 2001a,b; Gaona et al., 2011; Leuthardt et al., 2004; ing for ECoG, a more practical and robust platform Schalk et al., 2007). for BCI in clinical application. As detailed above, Besides providing more information content, because both EEG and single-unit-based systems have serious the ECoG signal is recorded from larger electrodes that The 46 BRIDGE do not penetrate the brain, ECoG sensors should have a strated that phonemic content taken from speech net- higher likelihood of long-term clinical durability. This works could be used for simple device control. expectation is supported by some pathologic and clini- Taken together, these studies show that ECoG sig- cal evidence. For example, in cat, dog, and monkey nals carry a high level of specific cortical information models, long-term subdural implants showed minimal and that these signals can enable a user to gain con- cortical or leptomeningeal tissue reaction while main- trol rapidly and effectively. It is worth noting, how- taining prolonged electrophysiologic recording (Bullara ever, that so far these control paradigms have not been et al., 1979; Chao et al., 2010; Loeb et al., 1977; Mar- extended to motor-impaired subjects. The effects on galit et al., 2003; Yuen et al., 1987). cortical signals in the setting of a spinal cord injury So far, ECoG for BCI applications has primarily been or amyotrophic lateral sclerosis (ALS) have not been studied in motor-intact patients with intractable epi- explicitly tested. lepsy that requires invasive monitoring. Preliminary work in humans using the implantable NeuroPace Conclusions device for the purpose of long-term subdural electrode The field of neuroprosthetics is growing rapidly as monitoring for the identification and abortion of sei- the cortical physiology that underpins the way a human zures has been shown to be stable (Vossler et al., 2004). brain encodes intentions is beginning to be understood. Similar to EEG-based BCI systems, the ECoG This understanding will have a significant impact in approach has primarily focused on changes in sensori­ improving function for people with various forms of motor rhythms from motor cortex. The major difference motor disability. As research moves beyond motor phys- has been access to higher frequency gamma rhythms iology, the field of neuroprosthetics is poised to increase with ECoG, which provide significant advantages in its capabilities and serve the needs of a more diverse training requirements and multidimensional control. clinical population. The evolving understanding of cor- In 2004, Leuthardt et al. demonstrated the first use tical physiology as it relates to motor movements, lan- of ECoG in closed-loop control in a one-dimensional guage function, and plasticity, could all provide higher cursor-control task with minimal training requirements levels of complexity in brain-derived control. (under 30 minutes). In additional experiments, the Given the rapid progression of these technologies over same group and others have demonstrated that specific the past decade and the concomitant rapid increase in frequency alterations encode very specific information computer processing speed, signal analysis techniques, about hand and arm movements (Leuthardt et al., 2004; and emerging ideas for novel biomaterials, we can be Pistohl et al., 2008; Sanchez et al., 2008; Schalk et al., hopeful that in the near future neuro­prosthetic implants 2007). In 2006, Leuthardt et al. (2006a) further dem- will be as common as deep brain stimulators are today. onstrated that ECoG control using a signal from the The clinical advent of this technology will usher in a epidural space was also possible. new era of restorative neurosurgery and new human- Schalk et al. (2008) has shown that ECoG signals can machine interfaces. be used for two-dimensional control with performance in the range of typical performance before the testing of Acknowledgments invasive single-unit systems. Because ECoG electrode The author acknowledges generous support from arrays cover broad regions of cortex, several groups the following: James S. McDonnell Foundation, have begun to explore alternate cognitive modalities Department of Defense (W911NF-07-1-0415 & and their cortical physiologies to expand BCI device W911NF-08-1-0216–Leuthardt), and the National control. Felton et al. (2007) have shown that, in addi- Institutes of Health (NINDS NIH R01-EB000856-06), tion to motor imagery, sensory imagery can also be used Children’s Discovery Institute. for device control. The same group demonstrated that auditory cortex could be trained to acquire simple con- Financial Disclosure trol of a cursor (Wilson et al., 2006). The author owns stock in the company Neurolutions. Ramsey et al. (2006) showed that higher cognitive functions, such as working memory in the dorsal lateral References prefrontal cortex, can also be used for effective device Abidian, M.R., and D.C. Martin. 2008. Experimental and operation. Recently, Leuthardt et al. (2011) demon- theoretical characterization of implantable neural micro- SPRING 2012 47

electrodes modified with conducting polymer nanotubes. Crone, N.E., D. Boatman, B. Gordon, L. and Hao. 2001b. Biomaterials 29(9): 1273–1283. Induced electrocorticographic gamma activity during audi- Andersen, R.A., J.W. Burdick, S. Musallam, B. Pesaran, and tory perception. Brazier Award-winning article. Clinical J.G. Cham. 2004. Cognitive neural prosthetics. Trends in Neurophysiology 112(4): 565–582. Cognitive Science 8(11): 486–493. Elbert, T., B. Rockstroh, W. Lutzenberger, and N. Birbaumer. Biran, R., D.C. Martin, and P.A. Tresco. 2005. Neuronal cell 1980. Biofeedback of slow cortical potentials. Electro- loss accompanies the brain tissue response to chronically encephalography and Clinical Neurophysiology 48(3): implanted silicon microelectrode arrays. Experimental 293–301. Neurology 195(1): 115–126. Farwell, L.A., and E. Donchin. 1988. Talking off the top of Birbaumer, N. 2006. Brain-computer-interface research: your head: toward a mental prosthesis utilizing event-relat- coming of age. Clinical Neurophysiology 117(3): 479–483. ed brain potentials. Electroencephalography and Clinical Birbaumer, N., N. Ghanayim, T. Hinterberger, I. Iversen, B. Neurophysiology 70(6): 510–523. Kotchoubey, A. Kübler, J. Perelmouter, E. Taub, and H. Felton, E.A., J.A. Wilson, J.C. Williams, and P.C. Garell. Flor. 1999. A spelling device for the paralysed. Nature 2007. Electrocorticographically controlled brain-comput- 398(6725): 297–298. er interfaces using motor and sensory imagery in patients Bjornsson, C.S., S.J. Oh, Y.A Al-Kofahi, Y.J. Lim, K.L. Smith, with temporary subdural electrode implants. Report of four J.N. Turner, S. De, B. Roysam, W. Shain, and S.J. Kim. cases. Journal of Neurosurgery 106(3): 495–500. 2006. Effects of insertion conditions on tissue strain and Freeman, W.J., M.D. Holmes, B.C. Burke, and S. Vanhatalo. vascular damage during neuroprosthetic device insertion. 2003. Spatial spectra of scalp EEG and EMG from awake Journal of Neural Engineering 3(3): 196–207. humans. Clinical Neurophysiology 114(6): 1053–1068. Blankertz, B., G. Dornhege, M. Krauledat, K.R. Muller, Gaona, C., M. Sharma, Z. Freudenburg, J. Breshears, D. V. Kunzmann, F. Losch, and G. Curio. 2006. The Ber- Bundy, J. Roland, D. Barbour, G. Schalk, and E. Leuthar- lin brain-computer interface: EEG-based communication dt. 2011. Nonuniform high-gamma (60–500 Hz) power without subject training. IEEE Transactions on Neural changes dissociate cognitive task and anatomy in human Systems and Rehabilitation Engineering 14(2): 147–152. cortex. Journal of Neuroscience 31(6): 2091–2100. Boulton, A.A., G.B. Baker, and C.H. Vanderwolf. 1990. Georgopoulos, A.P., J.F. Kalaska, R. Caminity, and J.T. Massey. Neurophysiological Techniques, II: Applications to Neural 1982. On the relations between the direction of two- Systems. New York: Humana Press. dimensional arm movements and cell discharge in primate Bullara, L.A., W.F. Agnew, T.G. Yuen, S. Jacques, and R.H. motor cortex. Journal of Neuroscience 2(11): 1527–1537. Pudenz. 1979. Evaluation of electrode array material for Georgopoulos, A.P., A.B. Schwartz, and R.E. Kettner. 1986. neural prostheses. Neurosurgery 5(6): 681–686. Neuronal population coding of movement direction. Sci- Carmena, J., M. Lebedev, R. Crist, J. O’Doherty, D. Santuc- ence 233(4771): 1416–1419. ci, D. Dimitrov, S. Patil, C. Henriquez, and M. Nicolelis. Heldman, D.A., W. Wang, S.S. Chan, D.W. and Moran. 2003. Learning to control a brain-machine interface for 2006. Local field potential spectral tuning in motor cor- reaching and grasping by primates. PLoS Biology 1(2): tex during reaching. IEEE Transactions on Neural System 193–208. Rehabilitation Engineering 14(2): 180–183. Chao, Z.C., Y. Nagasaka, and N. Fujii. 2010. Long-term Hochberg, L.R., M.D. Serruya, G.M. Friehs, J.A. Mukand, asynchronous decoding of arm motion using electrocorti- M. Saleh, A.H. Caplan, A. Branner, D. Chen, R.D. Penn, cographic signals in monkeys. Frontiers in Neuroengineer- and J.P. Donoghue. 2006. Neuronal ensemble control of ing 3: 3. prosthetic devices by a human with tetraplegia. Nature Crone, N.E., D.L. Miglioretti, B. Gordon, R.P. and Lesser. 442(7099): 164–171. 1998. Functional mapping of human sensorimotor cortex Huggins, J.E., S.P. Levine, S.L. BeMent, R.K. Kushwaha, L.A. with electrocorticographic spectral analysis. II. Event- Schuh, E.A. Passaro, M.M. Rohde, D.A. Ross, K.V. Elise- related synchronization in the gamma band. Brain 121 vich, and B.J. Smith. 1999. Detection of event-related (Pt 12): 2301–2315. potentials for development of a direct brain interface. Jour- Crone, N.E., L. Hao, J. Hart Jr., D. Boatman, R.P. Lesser, R. nal of Clinical Neurophysiology 16(5): 448–455. Irizarry, and B. Gordon. 2001a. Electrocorticographic Jarosiewicz, B., S.M. Chase, G.W. Fraser, M. Velliste, R.E. gamma activity during word production in spoken and sign Kass, and A.B. Schwartz. 2008. Functional network reor- language. Neurology 57(11): 2045–2053. ganization during learning in a brain-computer interface The 48 BRIDGE

paradigm. Proceedings of the National Academy of Sci- Rehabilitation Engineering 8(2): 180–185. ences of the U.S.A. 105(49): 19486–19491. Loeb, G.E., A.E. Walker, S. Uematsu, and B.W. Konigsmark. Kennedy, P.R., and R.A. Bakay. 1998. Restoration of neural 1977. Histological reaction to various conductive and output from a paralyzed patient by a direct brain connec- dielectric films chronically implanted in the subdural space. tion. Neuroreport 9(8): 1707–1711. Journal of Biomedical Materials Research 11(2): 195–210. Kübler, A., F. Nijboer, J. Mellinger, T.M. Vaughan, H. Pawel- Margalit, E., J.D. Weiland, R.E. Clatterbuck, G.Y. Fujii, M. zik, G. Schalk, D.J. McFarland, N. Birbaumer, and J.R. Maia, M. Tameesh, G. Torres, S.A. D’Anna, S. Desai, D.V. Wolpaw. 2005. Patients with ALS can use sensorimotor Piyathaisere, A. Olivi, E. de Juan Jr., and M.S. Humayun. rhythms to operate a brain-computer interface. Neurology 2003. Visual and electrical evoked response recorded from 64(10): 1775–1777. subdural electrodes implanted above the visual cortex in Laubach, M., J. Wessberg, and M.A. Nicolelis. 2000. Corti- normal dogs under two methods of anesthesia. Journal of cal ensemble activity increasingly predicts behaviour out- Neuroscience Methods 123(2): 129–137. comes during learning of a motor task. Nature 405(6786): McFarland, D.J., G.W., Neat, and J.R. Wolpaw. 1993. An 567–571. EEG-based method for graded cursor control. Psychobiol- Leuthardt, E.C., G. Schalk, J.R. Wolpaw, J.G. Ojemann, and ogy 21: 77–81. D.W. Moran. 2004. A brain-computer interface using McFarland, D.J., W. Sarnacki, and J.R. Wolpaw. 2008a. Elec- electrocorticographic signals in humans. Journal of Neural troencephalographic (EEG) control of three-dimensional Engineering 1(2): 63–71. movement. Society for Neuroscience Annual Meeting Leuthardt, E.C., K.J. Miller, G. Schalk, R.N. Rao, and J.G. 2008, Washington, D.C. Ojemann. 2005. Electrocorticography-Based Brain Com- McFarland, D., D. Krusienski, W. Sarnacki, and J. Wolpaw. puter Interface - The Seattle Experience. IEEE - Neural 2008b. Emulation of computer mouse control with a non- Systems and Rehabilitation Engineering 2005. invasive brain–computer interface. Journal of Neural Engi- Leuthardt, E.C., K.J. Miller, G. Schalk, R.P. Rao, and J.G. neering 5(2): 101–110. Ojemann. 2006a. Electrocorticography-based brain com- Millan, J.R., F. Renkens, J. Mourino, and W. Gerstner. 2004. puter interface—the Seattle experience. IEEE Transac- Noninvasive brain-actuated control of a mobile robot by tions on Neural Systems and Rehabilitation Engineering human EEG. IEEE Transactions on Biomedical Engineer- 14(2): 194–198. ing 51(6): 1026–1033. Leuthardt, E.C., G. Schalk, D. Moran, and J.G. Ojemann. Moran, D.W., and A.B. Schwartz. 1999a. Motor cortical 2006b. The emerging world of motor neuroprosthetics: a activity during drawing movements: population represen- neurosurgical perspective. Neurosurgery 59(1): 1–14; dis- tation during spiral tracing. Journal of Neurophysiology cussion 1–14. 82(5): 2693–2704. Leuthardt, E.C., G. Schalk, J. Roland, A. Rouse, and D.W. Moran, D.W., and A.B. Schwartz. 1999b. Motor cortical rep- Moran. 2009. Evolution of brain-computer interfaces: resentation of speed and direction during reaching. Journal going beyond classic motor physiology. Neurosurgical of Neurophysiology 82(5): 2676–2692. Focus 27(1): E4. Muller, K.R., M. Tangermann, G. Dornhege, M. Krauledat, G. Leuthardt, E.C., C. Gaona, M. Sharma, N. Szrama, J. Roland, Curio, and B. Blankertz. 2008. Machine learning for real- Z. Freudenberg, J. Solis, J. Breshears, and G. Schalk. 2011. time single-trial EEG-analysis: from brain-computer inter- Using the electrocorticographic speech network to control facing to mental state monitoring. Journal of Neuroscience a brain-computer interface in humans. Journal of Neural Methods 167(1): 82–90. Engineering 8(3): 036004. Nijboer, F., E.W. Sellers, J. Mellinger, M.A. Jordan, T. Matuz, Levine, S.P., J.E. Huggins, S.L. BeMent, R.K. Kushwaha, L.A. A. Furdea, S. Halder, U. Mochty, D.J. Krusienski, T.M. Schuh, E.A. Passaro, M.M. Rohde, and D.A. Ross. 1999. Vaughan, J.R. Wolpaw, N. Birbaumer, and A. Kubler. Identification of electrocorticogram patterns as the basis 2008. A P300-based brain-computer interface for people for a direct brain interface. Journal of Clinical Neurophysi- with amyotrophic lateral sclerosis. Clinical Neurophysiol- ology 16(5): 439–447. ogy 119(8): 1909–1916. Levine, S.P., J.E. Huggins, S.L. BeMent, R.K. Kushwaha, Pfurtscheller, G., D. Flotzinger, and J. Kalcher. 1993. Brain- L.A. Schuh, M.M. Rohde, E.A. Passaro, D.A. Ross, K.V. computer interface: a new communication device for hand- Elisevich, and B.J. Smith. 2000. A direct brain interface icapped persons. Journal of Microcomputer Applications based on event-related potentials. IEEE Transactions on 16(3): 293–299. SPRING 2012 49

Pfurtscheller, G., C. Guger, G. Mueller, G. Krausz, and C. in humans. Journal of Neural Engineering 4(3): 264–275. Neuper. 2000. Brain oscillations control hand orthosis in Schalk, G., K.J. Miller, N.R. Anderson, J.A. Wilson, M.D. a tetraplegic. Neuroscience Letters 292(3): 211–214. Smyth, J.G. Ojemann, D.W. Moran, J.R. Wolpaw, and E.C. Pfurtscheller, G., B. Graimann, J.E. Huggins, S.P. Levine, and Leuthardt. 2008. Two-dimensional movement control L.A. Schuh. 2003. Spatiotemporal patterns of beta desyn- using electrocorticographic signals in humans. Journal of chronization and gamma synchronization in corticographic Neural Engineering 5(1): 75–84. data during self-paced movement. Clinical Neurophysiol- Serruya, M.D., N.G. Hatsopoulos, L. Paninski, M.R. Fel- ogy 114(7): 1226–1236. lows, and J.P. Donoghue. 2002. Instant neural control of a Pistohl, T., T. Ball, A. Schulze-Bonhage, A. Aertsen, and C. movement signal. Nature 416(6877): 141–142. Mehring. 2008. Prediction of arm movement trajectories Seymour, J.P., and D.R. Kipke. 2007. Neural probe design for from ECoG-recordings in humans. Journal of Neurosci- reduced tissue encapsulation in CNS. Biomaterials 28(25): ence Methods 167(1): 105–114. 3594–3607. Polikov, V.S., P.A. Tresco, and W.M. Reichert. 2005. Response Spataro, L., J. Dilgen, S. Retterer, A.J. Spence, M. Isaacson, of brain tissue to chronically implanted neural electrodes. J.N. Turner, and W. Shain. 2005. Dexamethasone treat- Journal of Neuroscience Methods 148(1): 1–18. ment reduces astroglia responses to inserted neuropros- Ramsey, N.F., M.P. van de Heuvel, K.H. Kho, and F.S. Leijten. thetic devices in rat neocortex. Experimental Neurology 2006. Towards human BCI applications based on cognitive 194(2): 289–300. brain systems: an investigation of neural signals recorded Srinivasan, R., P.L. Nunez, and R.B. Silberstein. 1998. Spa- from the dorsolateral prefrontal cortex. IEEE Transac- tial filtering and neocortical dynamics: estimates of EEG tions on Neural System Rehabilitation Engineering 14(2): coherence. IEEE Transactions on Biomedical Engineering 214–217. 45(7): 814–826. Rohde, M.M., S.L. BeMent, J.E. Huggins, S.P. Levine, R.K. Sutter, E.E. 1992. The brain response interface: communi- Kushwaha, and L.A. Schuh. 2002. Quality estimation cation through visually-induced electrical brain responses. of subdurally recorded, event-related potentials based on Journal of Microcomputer Applications 15(1): 31–45. signal-to-noise ratio. IEEE Transactions on Biomedical Szarowski, D.H., M.D. Andersen, S. Retterer, A.J. Spence, Engineering 49(1): 31–40. M. Isaacson, H.G. Craighead, J.N. Turner, and W. Shain. Rouse, A.G., and D.W. Moran. 2009. Neural adaptation 2003. Brain responses to micro-machined silicon devices. of epidural electrocorticographic (EECoG) signals dur- Brain Research 983(1–2): 23–35. ing closed-loop brain computer interface (BCI) tasks. Pp. Taylor, D.M., S.I. Tillery, and A.B. Schwartz. 2002. Direct 5514–5517in Proceedings of the Annual International cortical control of 3D neuroprosthetic devices. Science Conference IEEE Engineering in Medicine and Biology 296(5574): 1829–1832. Society 2009. New York: IEEE. Vaughan, T.M., D.J. McFarland, G. Schalk, W.A. Sarnacki, Sanchez, J.C., A. Gunduz, P.R. Carney, and J.C. Principe. D.J. Krusienski, E.W. Sellers, and J.R. Wolpaw. 2006. The 2008. Extraction and localization of mesoscopic motor Wadsworth BCI Research and Development Program: at control signals for human ECoG neuroprosthetics. Journal home with BCI. IEEE Transactions on Neural Systems and of Neuroscience Methods 167(1): 63–81. Rehabilitation Engineering 14(2): 229–233. Schalk, G., D.J. McFarland, T. Hinterberger, N. Birbaumer, Velliste, M., S. Perel, M.C. Spalding, A.S. Whitford, and A.B. and J.R. Wolpaw. 2004a. BCI2000: a general-purpose Schwartz. 2008. Cortical control of a prosthetic arm for brain-computer interface (BCI) system. IEEE Transactions self-feeding. Nature 453(7198): 1098–1101. on Biomedical Engineering 51(6): 1034–1043. Vidal, J.J. 1977. Real-time detection of brain events in EEG. Schalk, G., E.C. Leuthardt, D. Moran, J. Ojemann, and J.R. IEEE Proceedings: Special Issue on Biological Signal Pro- Wolpaw. 2004b. Two-dimensional cursor control using cessing and Analysis 65: 633–664. electrocorticographic signals in humans. Presented at the Vossler, D.D., M. Doherty, R. Goodman, L. Hirsch, J. Young, Society for Neuroscience, San Diego, Calif., October 23, and D. Kraemer. 2004. Early safety experience with a fully 2004. implanted intracranial responsive neurostimulator for epi- Schalk, G., J. Kubanek, K.J. Miller, N.R. Anderson, E.C. Leu- lepsy. Presented at the Annual Meeting of the American thardt, J.G. Ojemann, D. Limbrick, D. Moran, L.A. Ger- Epilepsy Society, December 2004, New Orleans, La. hardt, and J.R. Wolpaw. 2007. Decoding two-dimensional Wang, W., S.S. Chan, D.A. Heldman, and D.W. Moran. movement trajectories using electrocorticographic signals 2007. Motor cortical representation of position and The 50 BRIDGE

velocity during reaching. Journal of Neurophysiology Wolpaw, J.R., and D.J. McFarland. 2004. Control of a two- 97(6): 4258–4270. dimensional movement signal by a noninvasive brain-com- Williams, J.C., J.A. Hippensteel, J. Dilgen, W. Shain, and puter interface in humans. Proceedings of the National D.R. Kipke. 2007. Complex impedance spectroscopy for Academy of Sciences of the U.S.A. 101(51): 17849–17854. monitoring tissue responses to inserted neural implants. Wolpaw, J.R., N. Birbaumer, W.J. Heetderks, D.J. McFarland, Journal of Neural Engineering 4(4): 410–423. P.H. Peckham, G. Schalk, E. Donchin, L.A. Quatrano, Wilson, J.A., E.A. Felton, P.C. Garell, G. Schalk, and J.C. C.J. Robinson, and T.M. Vaughan. 2000. Brain-computer Williams. 2006. ECoG factors underlying multimodal interface technology: a review of the first international control of a brain-computer interface. IEEE Transactions meeting. IEEE Transactions on Rehabilitation Engineer- on Neural Systems and Rehabilitation Engineering 14(2): ing 8(2): 164–173. 246–250. Wolpaw, J.R., N. Birbaumer, D.J. McFarland, G. Pfurtscheller, Wolpaw, J.R., D.J. McFarland, G.W. Neat, and C.A. Forneris. and T.M. Vaughan. 2002. Brain-computer interfaces for 1991. An EEG-based brain-computer interface for cursor communication and control. Clinical Neurophysiology control. Electroencephalography and Clinical Neurophys- 113(6): 767–791. iology 78(3): 252–259. Yuen, T.G., W.F. Agnew, and L.A. Bullara. 1987. Tissue Wolpaw, J.R., and D.J. McFarland. 1994. Multichannel EEG- response to potential neuroprosthetic materials implanted based brain-computer communication. Clinical Neuro- subdurally. Biomaterials 8(2): 138–141. physiology 90(6): 444–449. Researchers are working on implantable neurostimulators to treat debilitating neurological conditions, including blindness.

Retinal Prosthetic Systems for the Treatment of Blindness

James D. Weiland and Mark S. Humayun

Most functions of the human body are controlled by small electrical signals delivered via nerves. Thus, it is no surprise that electrical signals applied to the body from external sources can modulate physiological activ- James D. Weiland ity. In fact, reports of physiological electromodulation date back to the eighteenth century, but scientists of that time did not know enough about neurophysiology to understand the basic mechanism by which electricity could modulate biological activity. Today, with many advances in neuro­ biology, medicine, and engineering, we can inform the design of clinically beneficial, implantable neurostimulation devices to treat a number of debil- itating neurological diseases.

Neurostimulators Implantable neural stimulators activate nerve cells, which are respon- sible for processing and communicating information between the brain and other parts of the body (Kandel et al., 1991). Specialized nerve cells called Mark S. Humayun sensory receptors convert physical stimuli into electrical signals that can be relayed by other nerve cells to the brain. Because nerve cells are polarized,

James D. Weiland is Associate Professor of Ophthalmology and Biomedical Engineering, and Mark S. Humayun is Professor of Ophthalmology, Biomedical Engineering, and Cell and Neurobiology, and an NAE member. Both are at the Doheny Eye Institute, University of Southern California The 52 BRIDGE

an electrical potential can be measured across the cell Treating Blindness with Electrical Stimulation membrane. Transient changes in membrane poten- Causes of Blindness tial signal to other cells that an event has occurred, The retina is a light-sensitive, multilayer tissue that and neural networks, composed of connected neurons, lines the interior surface of the back of the eye (Figure 1) determine if that event, along with input from other (see http://webvision.med.utah.edu/). Photoreceptors cells, requires action by other parts of the nervous sys- (rods and cones) are the light-sensing cells of the tem. Events such as the onset of disease or an injury retina; the other cells process photoreceptor signals that damages nerve cells, particularly sensory cells, can and send information to the brain via the optic nerve. result in significant disability for the affected individ­ When photoreceptors degenerate due to disease, the ual, including loss of sensory input, diminished capabil- retina can no longer respond to light. However, suf- ity to process information, or reduced motor function. ficient numbers of other retinal nerve cells remain so How Electrostimulation Works that the electrical stimulation of these cells results in the perception of light. Activating a nerve cell by an electrical signal gener- Diseases such as retinitis pigmentosa and age-related ated by an implanted device is more complicated than macular degeneration cause blindness for millions of connecting two wires. The complexity arises partly people (Gehrs et al., 2010; Hartong et al., 2006) and from differences in carriers of the electrical charge. are presently untreatable. At first, symptoms are sub- In metals, electrons carry the charge, whereas in the tle, such as difficulty seeing at night or blurred central body, ions carry the charge. The conversion from elec- vision, but ultimately, these conditions result in blind- trons to ions occurs at the electrode, typically a metal ness. Given that vision is the sense by which people or metal oxide, in direct contact with the extracellu- obtain most of their information about their surround- lar fluid. An electrical signal applied to the electrode ings, blindness has an extremely detrimental impact on causes current to flow in the tissue via the movement the afflicted. of charged sodium, chloride, potassium, and other ions. The end effect is to depolarize the nerve cell mem- The Development of Electrical Stimulation brane. Depolarization beyond a certain level results in Electrical stimulation has been proposed as a treat- an action potential (as is the case with natural neural ment for blindness for decades, but only recently have signaling). An electrode has this affect on many neu- systems consistent with clinical use been developed. rons in its vicinity, and the summed activity of these The first documented use of electrical stimulation to neurons is an electrically elicited sensation or modu- create visual perception dates to 1755, when Charles lated function. LeRoy discharged a large capacitor through the head of Successful Neurostimulators A number of successful neurostimulators are in wide- spread use. For example, cochlear implants stimu- late the auditory nerve enabling deaf people to hear, often well enough to talk on a telephone. Implantable neurotransmitters­ can also relieve unrelenting pain that sometimes results from nerve damage or disease. For example, implantable devices that stimulate the lower spinal cord are known to decrease or even eliminate the feeling of pain. A dramatic example of electrical stimulation is in the treatment of Parkinson’s disease. By stimulating a part of the brain called the thalamus, many symptoms of Parkinson’s subside almost immediately. Parkinson’s FIGURE 1 Cross section of the retina and other layers of the eye (choroid and patients with uncontrollable tremor or rigidity that sclera). The rods and cones are the photoreceptors that sense light. The other cells severely limits motor function show improved coordi- process this information and transmit electrical impulses to the brain via optic nerve nation within minutes of commencing stimulation. fibers. Source: Image from http://www.catalase.com/retina.gif. SPRING 2012 53

a blind person, who described seeing “flames descending photons to digital data; a processing unit that generates downwards” (Marg, 1991). Over time, as science, medi- stimulus commands based on the image; analog drivers cine, and engineering have progressed, it has become that produce stimulus current; and an array of stimulat- feasible to develop a permanent implant to stimulate ing electrodes to deliver stimulus current to the retina the retina. (Weiland et al., 2005). Initial clinical experiments that provided proof-of- As shown in Figure 2, the electrode array can be principle involved briefly inserting a handheld electrode positioned in two locations in the eye, the epiretinal into the eye of a blind person, stimulating the retina, surface and the subretinal space. These anatomical and asking the person to describe what he or she saw, locations have come to define the two basic approaches if anything, and to describe the sensation (Humayun et being investigated for retinal prostheses. An epiretinal­ al., 1996; Rizzo et al., 2003). These simple yet essential implant would rest on the inner limiting membrane experiments established initial design parameters for a of the retina, whereas a subretinal implant would be permanent implantable device. inserted into the space occupied by photoreceptors in a healthy retina. Retinal Prostheses: General Description Several clinical trials have been conducted to test and Current Clinical Systems systems in blind humans. For the sake of brevity, only General Description the trials that have produced the most significant results A retinal prosthesis consists of several components are discussed below. that perform specific functions: a camera that converts

FIGURE 2 The retinal prosthesis concept. An image is captured by a camera and transmitted to an implant. The implant applies a patterned, complex stimulus to the retina via an array of electrodes on the surface of the retina (epiretinal) or underneath the retina (subretinal). Source: Image courtesy of Annual Review of Biomedical Engineering. The 54 BRIDGE

Subretinal Implant

Epiretinal Implant

FIGURE 3 (top) A subretinal implant developed by Retina Implant GmbH has been tested in 12 subjects. The microphotodiode array implanted in the subretinal space is to the far left. The remainder of the implant consists of cabling connected to external test equipment. (bottom) The epiretinal implant system developed by Second Sight Medical Products Inc., includes (left) a videocamera mounted in a pair of glasses, (center) a wearable image-processing system, and (right) an epiretinal implant with components inside and outside the eye. Stimulus commands generated by the wearable image-processing system are transmitted wirelessly to the implant via an inductive coil in the glasses frame. Source: Images courtesy of Proceedings of the Royal Society and the British Journal of Ophthalmology.

Subretinal Implant in 30 subjects (Figure 3, bottom). The device has 60 An externally powered, subretinal microphoto­diode electrodes and an external camera unit that delivers array (Figure 3 top), developed by Retinal Implant image information wirelessly to the implant. The best GmbH, has been tested in 12 subjects (Zrenner et al., result to date for visual acuity is 20/1,200 (Humayun et 2011). The device has 1,500 repeating units on a single al., 2011). silicon chip. Each unit has the following elements: a Twenty-two of the 30 subjects were able to read let- microphotodiode that senses light, digital and analog ters. However, with the ARGUS II device (as with the circuitry that scales a voltage stimulus based on the subretinal device described above), reading letters took sensed light, and a microelectrode that applies the stim- much longer than reading with natural vision. ARGUS ulus to the retina. The voltage stimulus pulse is supplied II subjects also demonstrated improved mobility, by a source outside the eye. could sense the direction of a moving object, and had The best test subject was able to read large letters improved hand-eye coordination (Ahuja et al., 2011). (although it took a long time to do so) and demon­ Assessing the Results strated visual acuity of approximately 20/1,000. Other subjects showed pattern recognition, light detection, The results described above for both devices have and object discrimination. generated considerable excitement among researchers in ophthalmology, vision, and biomedical engineer- EpiretinaI Implant ing. The possibility of restoring vision has captured the The ARGUS II retinal prosthesis (produced by Sec- imagination of many, and reports from the human test ond Sight Medical Products Inc.)1 has been implanted subjects testify to the promise of an implantable retinal prosthesis. In controlled tests, improved mobility was clearly evident, and subjects reported more confidence 1 Dr. Humayun has a financial interest in Second Sight Medical Products Inc. during ambulation. SPRING 2012 55

Nevertheless, based on an objective analysis of the face of the retina and account for anatomical variations data as a whole, one must conclude that we are a long among patients, each electrode will be able to activate a way from being able to claim “restored vision.” By all small part of the retina, thereby increasing visual acuity. measures, the tested individuals are still considered blind, even when the devices function well. The best Improving Stimulation and Rehabilitation reported visual acuity was 20/1,000, whereas the cutoff Better stimulation strategies will be necessary for per- for legal blindness is 20/200 (normal vision is 20/20). ceptions to appear more natural. Currently, perceptions In addition, the field of view is limited. The field for fade within seconds as neural adaptation mechanisms ARGUS II extends to 20 degrees (the limit for legal attenuate the artificial input. Researchers must focus blindness), but natural vision has a 180-degree field of on optimizing stimulus protocols to maximize user per- view. The subretinal results were even more limited, formance. Finally, rehabilitation strategies will be nec- with current implants providing a field of view of less essary to train users to maximize their performance with than 15 degrees. the implant.

Challenges for Artificial Vision The “Optogenetic” Approach Meeting the Basic Challenges A new approach to artificial vision could potentially We will need both technical and biological advances address some of the problems encountered with elec- to increase visual acuity with future implants. For tronic retinal prostheses. The “optogenetic” technique instance, simulations of artificial vision project that would modify individual neurons to incorporate light- about 1,000 individual pixels are necessary for visual sensitive ion channels into the cell membrane; the most function such as reading (at near normal speed) and face common light-sensitive channel is channelrhodopsin2 recognition. To reach this goal, we will need improve- (ChR2) (Gradinaru et al., 2010). Ion channels are the ments in electronic packaging (the materials and assem- means by which ions pass through the cell membrane, bly techniques that protect the circuits from saline). and membrane potential is influenced by whether chan- This will be particularly important for the subretinal nels are open or closed. When light of a specific wave- device described above, which now relies on a thin- length is shone on the cell, ChR2 ion channels open, film approach to protecting the subretinal electronics, resulting in depolarization of the cell. because a thick enclosure does not fit in the subretinal space. In addition, low-power integrated circuits must be developed that can both generate an effective stimu- A new, “optogenetic” lus pattern to evoke form perception and use power effi- ciently to ensure safety. approach to artificial vision

Connecting Electrodes to the Retina might address some of the Even if the barriers described above can be overcome, problems encountered with we need a better understanding of how to connect the device to the retina and how the brain will respond to bioelectronic prostheses. this type of input. The subretinal device described above has 1,500 elec- trodes, but it does not achieve the visual acuity that Bi et al. (2006) first demonstrated that this tech- would be possible if each electrode were acting indepen- nique could be used to modify retinal ganglion cells, dently. Higher visual acuity will require electrode arrays showing that light-evoked neural responses were pres- that are densely packed with small electrodes. How- ent in a mouse model of retinal degeneration when the ever, to take advantage of this density, the electrode mouse retinal cells contained ChR2. Others have since arrays must be consistently positioned in close contact expanded on this work. with the retina. Advantages Inconsistent positioning of the electrode array has been a major barrier for epiretinal implants. If electrode The optogenetic approach has some significant arrays can be developed that can conform to the sur- advantages over the bioelectronic approach. By making The 56 BRIDGE each cell light sensitive, vision can potentially be Although important breakthroughs have been made, restored to near-normal acuity. Also, by using light as this field is still in the early stages. Moving toward the the activating signal, the optics of the eye can focus an restoration of high-acuity vision will require concerted image on the retina. In other words, the optogenetic efforts by scientists, engineers, clinicians, and, most technique can come much closer to restoring natural important, blind patients. vision than bioelectronic approaches. References Disadvantages Ahuja, A.K., J.D. Dorn, A. Caspi, M.J. McMahon, G. Dagne- Artificial vision based on optogenetics also has chal- lie, L. daCruz, P Stanga, M.S. Humayun, R.J. Greenberg; lenges that preclude clinical use. The main issue relates Argus II Study Group. 2011. Blind subjects implanted to sensitivity. with the Argus II retinal prosthesis are able to improve per- Currently, modified cells require that bright blue formance in a spatial-motor task. British Journal of Oph- light (460 nm) be activated, roughly 7 orders of mag- thalmology 95(4): 539–543. doi:10.1136/bjo.2010.179622. nitude above the light-sensitivity threshold of normally Bi, A., J. Cui, Y.-P. Ma, E. Olshevskaya, M. Pu, A.M. Dizhoor, sighted people. Thus, an external apparatus would still and Z.-H. Pan. 2006. Ectopic expression of a microbial- be required to convert an image from a camera into a type rhodopsin restores visual responses in mice with pho- light stimulus that interacts with the modified cells. In toreceptor degeneration. Neuron 50(1): 23–33. addition, it is not known if cells can be modified perma- Gehrs, K.M., J.R. Jackson, E.N. Brown, R. Allikmets, and nently, or if repeated injections would be needed. G.S. Hageman. 2010. Complement, age-related macular Finally, it is not clear how such intense light would degneration and a vision of the future. Archives of Oph- interact with a diseased retina, which has only rem- thalmology 128(3): 349–358. nant light sensitivity. A common symptom of retinitis Gradinaru, V., F. Zhang, C. Ramakrishnan, J. Mattis, R. pigmentosa, for example, is photophobia (discom- Prakash, I. Diester, I. Goshen, K.R. Thompson, and K. fort in bright light), which is clearly not compatible Deisseroth. 2010. Molecular and cellular approaches for with a therapy that requires intense light input into diversifying and extending optogenetics. Cell 141(1): the eye. 154–165. Hartong, D.T., E.L. Berson, and T.P. Dryja. 2006. Retinitis Summary pigmentosa. The Lancet 368(9549): 1795–1809. These are interesting times for researchers on retinal Humayun, M.S., E. de Juan Jr., G. Dagnelie, R.J. Greenberg, prostheses. Clinical trials have shown both the prom- R.H. Propst, and D.H. Phillips. 1996. Visual perception ise and limitations of electrical stimulation as a treat- elicited by electrical stimulation of retina in blind humans. ment for blindness. People working in this field are Archives of Ophthalmology 114(1): 40–46. constantly reminded of the wondrous sense of vision Humayun, M.S., L. da Cruz, G. Dagnelie, J.A. Sahel, P.E. available to most of us, how reliant we are on our vision, Stanga, E. Filley, D. Eliott, J.L. Duncan, and R.J. Green- and the devastating impact of vision loss. Given the berg. 2011. Interim performance results from the Second complexity of vision, for the foreseeable future, pros- Sight Argus II retinal prosthesis study. Investigative Oph- thetic vision systems will clearly provide vision that is thalmology & Visual Science 52:E-Abstract 2594. artificial in appearance and below the resolution neces- Kandel, E.R., J.H. Schwartz, and T.M. Jessel. 1991. Principles sary for complex, visually guided tasks. of Neural Science, 3rd edition. Norwalk, Conn.: Apple- However, even a slight improvement in vision can ton & Lange. have a significant impact on quality of life. Restor- Marg, E. 1991. Magnetostimulation of vision: direct non- ing a person’s ability to see large objects and detect invasive stimulation of the retina and the visual brain. motion can increase confidence for navigating through Optometry and Vision Science 68(6): 427–440. unfamiliar environments. In addition, the brain has Rizzo, J.F. 3rd, J. Wyatt, J. Loewenstein, S. Kelly, and D. Shire. an amazing ability to adapt to new input, and an indi- 2003. Methods and perceptual thresholds for short-term vidual, through experience and training, can use other electrical stimulation of human retina with microelectrode contextual and sensory information to improve his or arrays. Investigative Ophthalmology & Visual Science her understanding of what is being “seen” via an artifi- 44(12): 5355–5361. cial vision system. Weiland, J.D., W. Liu, and M.S. Humayun. 2005. Retinal SPRING 2012 57

prosthesis. Annual Review of Biomedical Engineering Stingl, H. Sachs, A. Stett, P. Szurman, B. Wilhelm, and 7:361–401. R. Wilke. 2011. Subretinal electronic chips allow blind Zrenner, E., K.U. Bartz-Schmidt, H. Benav, D. Besch, A. patients to read letters and combine them to words. Pro- Bruckmann, V.P. Gabel, F. Gekeler, U. Greppmaier, A. ceedings of the Royal Society of London, Series B: Biologi- Harscher, S. Kibbel, J. Koch, A. Kusnyerik, T. Peters, K. cal Sciences 278(1711): 1489–1497. The 58 BRIDGE NAE News and Notes Class of 2012 Elected

In February, NAE elected 66 Craig H. Benson, Wisconsin Dis- For establishing engineering educa- new members and 10 new foreign tinguished Professor, director of sus- tional programs for nuclear reactor associates, bringing the number tainability research and education, design and safety. of U.S. members to 2,254 and the and chair of Civil and Environmen- George M. Church, director, number of foreign associates to tal Engineering and of Geological Center for Computational Genet- 206. Engineers are elected to NAE Engineering, University of Wiscon- ics, and professor of genetics, Har- for outstanding contributions to sin, Madison. For improvements in vard Medical School, Boston. For “engineering research, practice, or design, construction, and monitor- contributions to human genome education, including . . . signifi- ing of earthen liners and covers for sequencing technologies and DNA cant contributions to the engineer- municipal, hazardous, and radio­ synthesis and assembly. ing literature” and to “new and active waste landfills. Jared L. Cohon, president and developing fields of technology, . . . Barbara D. Boyan, professor professor of civil and environmen- major advancements in traditional and Price Gilbert Jr. Chair in Tis- tal engineering, Carnegie Mellon fields of engineering, or . . . inno- sue Engineering; associate dean for University, Pittsburgh. For contri- vative approaches to engineering research, College of Engineering; butions to environmental systems education.” A list of newly elected and Georgia Research Alliance analysis and national policy and members and foreign associates Eminent Scholar, Wallace H. Coul- leadership in higher education. follows, with primary affiliations ter Department of Biomedical Engi- James J. Coleman, Intel Alumni at the time of their election and neering, Institute for Bioengineering­ Endowed Chair in Electrical and brief descriptions of their principal and Bioscience, Georgia Institute of Computer Engineering, professor of accomplishments. Technology, Atlanta. For engineer- materials science and engineering, Mark Adamiak, director of ing implant technologies for bone and and director of the Semiconductor advanced technologies, GE Digital cartilage repair. Laser Laboratory, University of Illi- Energy Multilin, Wayne, Pennsyl- Mary C. Boyce, Ford Profes- nois, Urbana-Champaign. For con- vania. For contributions to power sor of Engineering and department tributions to semiconductor lasers system protection, control, moni- head of Mechanical Engineering, and photonic materials. toring, and communications. Massachusetts Institute of Technol- Louis Anthony (Tony) Cox Jr., Robert D. Allen, senior man- ogy, Cambridge. For contributions president, Cox Associates, Denver, ager, Advanced Materials Chem- to understanding the mechanics of Colorado. For applications of oper- istry Department, IBM Almaden deformation in engineered and nat- ations research and risk analysis to Research Center, San Jose, Califor- ural polymeric solids. significant national problems. nia. For innovations in chemistries Joan F. Brennecke, Keating- Robert L. Crippen, former astro- and materials for semiconductor Crawford Professor of Chemical naut and director of the NASA manufacturing. and Biomolecular Engineering and Kennedy Space Center; and retired Michael I. Baskes, adjunct pro- director of the Energy Center, Uni- president, Thiokol Propulsion fessor, Department of Mechanical versity of Notre Dame, Notre Dame, Group, Palm Beach Gardens, Flori- and Aerospace Engineering, Jacobs Indiana. For innovation in the use da. For leadership in human space School of Engineering, University of ionic liquids and supercritical flight and development of solid of California, San Diego, La Jolla. fluids for environmentally benign fueled rockets. For contributions to the embedded chemical processing. Supriyo Datta, Thomas Duncan atom method for predicting the Max William Carbon, Professor Distinguished Professor of Electrical structure and properties of metals Emeritus of Nuclear Engineering, and Computer Engineering, Purdue and alloys. University of Wisconsin, Madison. University, West Lafayette, Indiana. SPRING 2012 59

For quantum transport modeling in ResMed Inc., San Diego, Califor- Center, Yorktown Heights, New nanoscale electronic devices. nia. For research and development York. For contributions to low Akhil Datta-Gupta,­ Regents Pro- of devices for treatment of sleep dis- dielectric constant insulators for fessor and L.F. Peterson ’36 Chair, order breathing. the interconnects of very large Harold Vance Department of Petro- James R. Fienup, Robert E. Hop- scale integrated electronic devices. leum Engineering, Texas A&M kins Professor of Optics; professor, Victoria F. Haynes, president and University, College Station. For Center for Visual Science; senior chief executive officer, RTI Interna- developing the theory and practice scientist, Laboratory for Laser Ener- tional, Research Triangle Park, North of streamline simulation for fluid getics; and professor of electrical and Carolina. For integrating research, flow in heterogeneous reservoirs. computer engineering; University economics, and social sciences to William P. Delaney, Director’s of Rochester, Rochester, New York. solve multidisciplinary problems. Office Fellow, MIT Lincoln Labo- For development and applications Richard Hogg, Professor Emeri- ratory, Lexington, Massachusetts. of phase retrieval algorithms. tus of Mineral Processing and For contributions to radar systems Huajian Gao, Walter H. Annen- Geoenvironmental Engineering, for national defense. berg Professor of Engineering, Pennsylvania State University, Steven P. DenBaars, Mitsubishi School of Engineering, Brown Uni- University Park. For contributions Chemical Professor in Solid State versity, Providence, Rhode Island. to the science and engineering of Lighting and Displays, Materials For contributions to micromechan- coagulation and flocculation in par- Department, University of Cali- ics of thin films and hierarchically ticulate systems. fornia, Santa Barbara. For contri- structured materials. Ray R. Irani, executive chair- butions to gallium nitride-based William W. George, professor man, Occidental Petroleum Cor- materials and devices for solid state of management practice, Harvard poration, Los Angeles, California. lighting and displays. Business School, Cambridge, Mas- For leadership in the petrochemical Dennis E. Discher, Robert D. sachusetts. For applying engineer- industry and processes for applica- Bent Professor of Chemical and ing principles to manufacturing to tions of particulate systems. Biomolecular Engineering, Univer- advance health care. James W. Jones, Distinguished sity of Pennsylvania, Philadelphia. Peter W. Glynn, Thomas W. Professor, Department of Agricul- For elucidation of the effects of Ford Professor and chair, Manage- tural and Biological Engineering, mechanical forces on cell physiol- ment Science and Engineering University of Florida, Gainesville. ogy and stem cell development. Department, Stanford University, For contributions to understand- Elazer R. Edelman, Thomas D. Stanford, California. For contri- ing climate change, environmental and Virginia W. Cabot Professor of butions to simulation methodology impacts, and sustainable agricul- Health Sciences and Technology, and stochastic modeling. tural systems. Massachusetts Institute of Technol- Alan H. Gnauck, Distinguished Mujid S. Kazimi, TEPCO Pro- ogy, Cambridge. For contributions Member of the Technical Staff, Bell fessor of Nuclear Engineering, to the design, development, and reg- Labs, Alcatel-Lucent, Holmdel, professor of mechanical engineer- ulation of local cardiovascular drug New Jersey. For contributions to ing, and director of the Center for delivery and drug eluting stents. high-speed, high-capacity lightwave Advanced Nuclear Energy Systems, Gordon R. England, president, E6 communications systems. Massachusetts Institute Technol- Partners LLC, Arlington, Virginia. Steven M. Gorelick, Cyrus F. ogy, Cambridge. For contributions For advances in digital avionics for Tolman Professor, Department of to technologies for the nuclear fuel aircraft, land, and naval platforms. Environmental Earth System Sci- cycle and reactor safety. Robert W. Farquhar, executive ence, Stanford University, Stan- Richard Wilker Korsmeyer, for space exploration, KinetX Inc., ford, California. For optimization senior research fellow and head of Tempe, Arizona. For deep space mis- techniques and transport models for business development and licensing sions to asteroids and comets and for groundwater and the remediation of for pharmaceutical sciences, Pfizer leading the NEAR mission to Eros. contaminated aquifers. Inc., New London, Connecticut. Peter C. Farrell, founder, chair- Alfred Grill, IBM Fellow, For contributions to drug delivery man, and chief executive officer, IBM Thomas J. Watson Research formulations and medical devices. The 60 BRIDGE

Helmut Krawinkler, John A. Jyotirmoy Mazumder, Robert H. Philip M. Neches, chairman, Blume Professor Emeritus of Engi- Lurie Professor of Mechanical Engi- Foundation Ventures LLC, New neering, Stanford University, Stan- neering, professor of materials sci- York City. For the architecture ford, California. For development ence and engineering, director of the and software of parallel database of performance-based earthquake Center for Laser-Aided Intelligent appliances. engineering procedures for evaluat- Manufacturing, and director of the James J. O’Brien, consulting engi- ing and rehabilitating buildings. NSF I/UCRC for Lasers and Plasmas neer, James J. O’Brien P.E., Riverton, Fikri J. Kuchuk, fellow and for Advanced Manufacturing, Uni- New Jersey. For development of chief reservoir engineer, Schlum- versity of Michigan, Ann Arbor. For standards of practice for computer- berger Testing Services, Clamart, quantitative transport modeling for ized scheduling of construction proj- France. For contributions in pres- laser interaction and for the design ects and capital programs. sure transient analyses for petro- and commercialization of direct Babatunde A. Ogunnaike, interim leum reservoirs. metal deposition machines. dean; William L. Friend Chair of Juan C. Lasheras, Stanford S. and Diane M. McKnight, fellow of Chemical Engineering; and profes- Beverley P. Penner Distinguished the Institute of Arctic and Alpine sor, Center for Systems Biology-DBI; Chair Professor of Engineering and Research and professor of civil, University of Delaware, Newark. Applied Sciences, Distinguished environmental and architectural For advances in process systems, Professor of Engineering, and direc- engineering, University of Colo- process engineering practice, and tor of the Center for Medical Devices rado, Boulder. For elucidating the systems engineering education. and Instrumentation, Institute of interrelationship between natural Norbert Joseph Pelc, profes- Engineering in Medicine, Universi- organic matter and heavy metals in sor and associate chair for research ty of California, San Diego, La Jolla. streams and lakes. of the Radiology Department, and For studies of atomization, turbulent Antonios Georgios Mikos, Louis professor of electrical engineering mixing, and heat transfer and for the Calder Professor of Bioengineering (by courtesy), Stanford University, development of medical devices. and Chemical and Biomolecular Stanford, California. For develop- Kai Li, Paul M. Wythes ’55, P’86 Engineering; director, J.W. Cox Lab- ment of algorithms and technolo- and Marcia R. Wythes P’86 Pro- oratory for Biomedical Engineering; gies for MRI, CT, and hybrid X-ray/ fessor, Department of Computer and director, Center for Excellence MRI imaging. Science, Princeton University, in Tissue Engineering; Rice Univer- Leonard Pinchuk, co-founder, Princeton, New Jersey. For advances sity, Houston, Texas. For advances president, and chief executive in data storage and distributed com- in tissue engineering, regenerative officer, Innovia LLC and Related puter systems. medicine, biomaterials, and drug Companies, Miami, Florida. For Henrique S. Malvar, chief sci- delivery, including development of development of biomedical poly- entist and Distinguished Engineer, biodegradable polymers. meric materials for angioplasty Microsoft Research, Redmond, Richard K. Miller, president and balloons, drug eluting stents, and Washington. For contributions to professor of mechanical engineer- other devices. multiresolution signal processing ing, Franklin W. Olin College of Andrea Prosperetti, Charles A. and multimedia signal compression Engineering, Needham, Massachu- Miller Jr. Professor of Mechanical and standards. setts. For establishing a new para- Engineering, Johns Hopkins Uni- Tobin J. Marks, Charles E. and digm for undergraduate engineering versity, Baltimore, Maryland. For Emma H. Morrison Professor of education and for the establishment contributions to the fundamentals Chemistry, professor of materi- of Olin College. and applications of multiphase als science and engineering, and Robert A.K. Mitchell, consul- flows. Vladimir N. Ipatieff Professor of tant and retired vice president, Kaushik Rajashekara, chief Catalytic Chemistry, Northwestern Northrop Grumman Aerospace Sys- technologist, Electric Power and University,­ Evanston, Illinois. For tems, Redondo Beach, California. Control Systems, Rolls-Royce Cor- innovation in electronic, photonic, For development of autonomous poration, Indianapolis, Indiana. For and photovoltaic materials and cat- unmanned aerial systems and their contributions to electric power con- alytic polymerization. applications. version systems in transportation. SPRING 2012 61

Nambirajan Seshadri, senior Rogers H. Stolen, Distinguished Materials Science and Technology vice president and general man- Visiting Professor in Materials Sci- Division, Oak Ridge National Labo- ager, Mobile Wireless Group, and ence and Engineering and faculty ratory, Oak Ridge, Tennessee. For chief technology officer, Mobile member in the Center for Optical advancing understanding of radia- Platforms and Wireless Connectiv- Materials Science and Engineering tion damage in metallic and ceramic ity, Broadcom Corporation, Irvine, Technologies, Clemson University, components. California. For contributions to Clemson, South Carolina. For con- wireless communications theory tributions to fiber nonlinear optics New Foreign Associates and development of mass market and invention of polarization pre- E. (Edward) John Hinch, profes- wireless technology. serving fiber. sor of fluid mechanics, Department of David E. Shaw, chief scientist, Samuel I. Stupp, director, Insti- Applied Mathematics and Theoreti- D.E. Shaw Research, New York tute for Bionanotechnology in cal Physics, Centre for Mathematical City. For the architecture, design, Medicine; and Board of Trustees Sciences, University of Cambridge, and implementation of the Anton Professor of Materials Science, Cambridge, United Kingdom. For protein-folding supercomputer. Chemistry, and Medicine; North- contributions to the mechanics of Scott J. Shenker, professor, Elec- western University, Evanston, Illi- fluids, suspensions, and polymeric trical Engineering and Computer nois. For advances in processes of liquids and to industrial processes. Science Department, University self-assembled polymers for biomed- Sue E. Ion, consultant and vis- of California, Berkeley. For con- ical applications. iting professor, Imperial College- tributions to Internet design and Gordana Vunjak-Novakovic, London, Leyland, United Kingdom. architecture. vice chair and professor of biomedi- For contributions to nuclear fuel Christine A. Shoemaker, Joseph cal engineering and director of the development. P. Ripley Professor of Engineering, Laboratory for Stem Cells and Tis- Frank P. Kelly, professor of the School of Civil and Environmental sue Engineering, Columbia Univer- mathematics of systems and mas- Engineering, Cornell University, sity, New York City. For bioreactor ter of Christ’s College, University Ithaca, New York. For develop- systems and modeling approaches of Cambridge, Cambridge, United ment of decision-making optimiza- for tissue engineering and regenera- Kingdom. For contributions to the tion algorithms for environmental tive medicine. theory and optimization of commu- and water resources problems. Michael S. Waterman, University nication networks. Amit Singhal, Google Fellow, Professor, USC Associates Chair in Kinam Kim, president, Samsung Google Inc., Mountain View, Cali- Natural Sciences, and professor of Electronics Company, Yongin-si, fornia. For contributions to infor- biological sciences, computer sci- Gyeonggi-do, Korea. For contribu- mation retrieval and search. ence, and mathematics, College of tions to semiconductor technolo- Robert E. Skelton, Daniel L. Letters, Arts, and Sciences, Uni- gies for DRAM and nonvolatile Alspach Professor of Dynamic versity of Southern California, Los memories. Systems and Controls (Emeritus) Angeles. For development of com- Chao-Han Liu, Distinguished and director, Structural Systems putational methods for DNA and Visiting Scholar, Academia Sinica, and Control Laboratory, Univer- protein sequence analyses. Taipei, Taiwan. For contributions sity of California, San Diego, La K. Dane Wittrup, C.P. Dubbs Pro- to ionospheric research and inter- Jolla. For contributions to robust fessor of Chemical Engineering and national leadership in atmospheric control, system identification, and Biological Engineering, and associate remote sensing. methodology for control-structure director, Koch Institute for Integra- Gennady A. Mesyats, director, interaction. tive Cancer Research, Massachusetts P.N. Lebedev Physical Institute and David A. Stahl, professor, Institute of Technology, Cambridge. vice president, Russian Academy Department of Civil and Environ- For developments in protein engi- of Sciences, Moscow. For devel- mental Engineering, University of neering, protein expression, and opment and application of pulsed Washington, Seattle. For applica- quantitative pharmacology. power technology. tion of molecular microbial ecology Steven J. Zinkle, UT-Battelle Pradip, chief scientist and head to environmental engineering. Corporate Fellow and director, of the Process Engineering Lab, Tata The 62 BRIDGE

Research Development and Design cing the understanding of defor- and development of protein thera- Centre, Pune, India. For contribu- mation and fracture of structural peutic platforms. tions to processing of minerals and materials and development of mate- Andrés Weintraub, professor, waste materials. rials research infrastructure for soci- Department of Industrial Engineer- P. Rama Rao, chairman, Inter- etal needs. ing, University of Chile, Santiago. national Advanced Research Cen- Willem P.C. Stemmer, chief For deployment of innovative tre for Powder Metallurgy and New executive officer, Amunix Inc., decision support systems for natu- Materials (ARCI), Hyderabad, Mountain View, California. For ral and human resources in South Andhra Pradesh, India. For advan­ co-invention of directed evolution America.

NAE Newsmakers

John E. Bowers, director, Insti- Magazine and industry experts based Joseph DeSimone, Chancellor’s tute for Energy Efficiency, and pro- on nominations from people around Eminent Professor of Chemistry, fessor, Department of Electrical and the world. Dr. Chopra was selected University of North Carolina, and Computer Engineering, University for “pioneering methods for earth- William R. Kenan Jr., Distinguished of California, Santa Barbara, is the quake response analysis of structural Professor of Chemical Engineering, recipient of the 2012 John Tyn- systems, including dynamic inter­ North Carolina State University, will dall Award. Professor Bowers was action with soils and fluids” and for receive Sigma Xi’s 2012 Walston selected for international leader- “encouraging students to study very Chubb Award for Innovation. Dr. ship in the development of novel tough issues associated with con- DeSimone has published more than optoelectronic devices, including crete dams (gravity and arch) since 260 scientific articles, has been issued “groundbreaking research in hybrid- the 1980s.” Dr. Chopra is also a more than 115 patents in his name, silicon lasers and photonic integrated consultant on earthquake engineer- and has another 120 patents pend- circuits.” This hybrid technology ing to numerous government and ing. The Chubb Award, which has lowers the cost of photonic sub- private organizations. been given since 2006 to honor and systems, making it practical to apply The Industrial Research Insti- promote creativity in science and optical communication technol- tute presented the 2011 IRI Medal engineering, includes a $4,000 hono- ogy in areas where it has been pro- to Uma Chowdhry, senior vice rarium and an invitation to deliver hibitively expensive. The Tyndall president and chief science and the Walston Chubb Award Lecture Award is presented by the Optical technology officer at DuPont from at Sigma Xi’s annual meeting. Society and the IEEE Photonics 2006 until her retirement in 2010. Solomon W. Golomb, Andrew Society. Professor Bowers received She was honored for “champion- and Erna Viterbi Professor of Com- the award in Los Angeles during the ing innovations to advance envi- munications and Distinguished plenary session of the 2012 Opti- ronmentally sustainable solutions; University Professor, University cal Fiber Communication Confer- building robust, multidisciplinary of Southern California, has been ence and Exposition/National Fiber collaborations between R&D, selected as the recipient of Sigma Optic Engineers, March 4–8. marketing, and manufacturing to Xi’s 2012 William Procter Prize Anil K. Chopra, Johnson Pro- deliver competitively differentiated, for Scientific Achievement. Dr. fessor of Structural Engineering, customer-focused solutions world- Golomb is being honored for playing University of California Berkeley, wide; developing strategic partner- a key role in formulating the design has been included in the list of ships with government, universities, of deep-space communications for “20 people who have made the and corporations to harness the lunar and planetary exploration, biggest difference in dam engi- creativity of the entire innovation while he was assistant chief of the neering over the last 10 years.” ecosystem; championing the role of Telecommunications Research Sec- Individuals on the list are selected women in science and engineering; tion at the Jet Propulsion Labora- by the editorial team of International and mentoring science and engi- tory. The Procter Prize has been Water Power and Dam Construction neering professionals.” presented annually since 1950. SPRING 2012 63

John M. Kulicki, chairman and related to the impact of atmospheric “corresponding member” in the Class CEO of the bridge engineering and space processes on terrestrial of Physical Sciences by the presti- firm Modjeski and Masters, is the and space technologies.” The pre- gious Bologna Academy of Sciences. recipient of the prestigious J. Lloyd sentation was made at the Honors The Global Semiconductor Alli- Kimbrough Award “for significant Ceremony at the 2011 AGU fall ance (GSA) has announced that achievements in bridge analysis and meeting in San Francisco on Decem- Henry Samueli, industry pioneer design.” The award is presented ber 7, 2011. The Bowie Medal was and co-founder and chief technol- by the American Institute of Steel established in 1939 in honor of ogy officer, Broadcom Corporation, Construction to honor pre-eminent William Bowie for his “spirit of help- is the winner of the 2011 Dr. Mor- engineers and architects who have fulness and friendliness in unselfish ris Chang Exemplary Leadership made outstanding contributions to cooperative research.” Award. Dr. Samueli was presented the structural steel industry through Cleve B. Moler, chairman and with this lifetime achievement award their design work. Dr. Kulicki Chief Scientist of The MathWorks, during the GSA Awards Dinner Cel- will receive the award at the 2012 Inc., received the IEEE Computer ebration on December 8, 2011. NASCC (North American Steel Society Sidney Fernbach Award. The Association for Computing Construction Conference): The Dr. Moler received a certificate and Machinery (ACM) and the IEEE Steel Conference, which will be a $2,000 honorarium for “fundamen- Computer Society has awarded the held in Dallas on April 18–21, 2012. tal contributions to linear algebra, 2011 Seymour Cray Computer Louis J. Lanzerotti, Distinguished mathematical software, and enabling Engineering Award to Charles L. Research Professor of Physics, New tools for computational science”—in Seitz, researcher, writer, and con- Jersey Institute of Technology, November 2011 in Seattle at SC11, sultant, and a founder of Myricom, was selected as the 2011 William an international conference of high- Inc. The award is given in recogni- Bowie Medalist by the American performance computing, network- tion of innovative contributions to Geophysical Union (AGU). Dr. ing, storage, and analysis. high-performance computing sys- Lanzerotti was honored for more Shlomo P. Neuman, Regents’ tems that exemplify Seymour Cray’s than 40 years of contributions “to Professor, Department of Hydrol- creative spirit. The winner receives research on space plasmas and geo- ogy and Water Resources, Univer- a crystal memento, a certificate, and physics and engineering problems sity of Arizona, has been elected a a $10,000 honorarium.

Changing the Conversation Is Now on Facebook

The 2008 report, Changing the and received support from NSF to (www.facebook.com/engineersctc) Conversation: Messages for Improv- create an online messaging “tool- intended to build awareness and ing Public Understanding of Engineer- kit” (www.enginereingmessages.org), increase the adoption of CTC mes- ing, presented a new positioning which was launched one year ago. sages by new audiences, such as statement for engineering and some The site provides background on engineering administrators, fac- new, research-based messages and the engineering messaging “prob- ulty, and students; leaders in pro- taglines. The new messages have lem” and tips for using messages, fessional engineering societies; and been adopted by a number of engi- catalogs nearly 130 examples of leaders and mid-level managers in neering professional societies and messaging about engineering and technology-intensive industries, a few schools of engineering, but case studies of successful messaging, particularly those responsible for by hardly any technology-focused and includes interactive features communication, marketing, and hir- companies. that enable members to share and ing. The UEF-funded project will To make the Changing the Con- discuss their work. also promote the CTC messages and versation (CTC) report more With support from the United website through Twitter and LinkedIn widely available to the engineer- Engineering Foundation (UEF), NAE and will monitor and interact with ing community, NAE applied for recently launched a Facebook page relevant blogs. The 64 BRIDGE

New Mirzayan Science and Technology Policy Graduate Fellows

demic training and internships, she has conducted biomedical research on a wide range of subjects, includ- ing tissue engineering, lab-on-chip microdevices, and neurobiology. At Penn, Mahlet chaired the Black Graduate and Professional Student Assembly and co-founded the Penn Graduate Women in Sci- ence and Engineering Program with Kristen Coakley Mahlet Mesfin five other students. Through her work on diversity with these groups, Kristen Coakley received her a volunteer, she discovered a love she not only discovered a love for Ph.D. in biomedical sciences from of conveying the implications of policy and advocacy, but also devel- the University of California, San scientific research to the general oped skills for addressing change on Francisco. She also holds a bach- public and an interest in making a university-wide level She hopes elor’s degree in biological sciences science education more effective. her Mirzayan Fellowship will pro- from Mount Holyoke College. As a At NAE, Kristen is working vide hands-on experience that research technician and Ph.D. can- on the Integrated STEM Project. merges her scientific interests with didate, her interests were focused on When she is not working, she loves a deeper understanding of policy at the development and function of the to shop at farmers’ markets and cook the federal level. adaptive immune system, including the tasty ingredients she finds there. During her internship, Mahlet is mechanisms underlying autoimmu- working with the NAE Program on nity and lymphomagenesis. Mahlet Mesfin recently com- Diversity in the Engineering Work- While pursuing her Ph.D., Kris- pleted her Ph.D. in bioengineering force and the Committee on Women ten designed and taught science at the University of Pennsylvania. in Science, Engineering, and Math- lessons for grades 1–6 in the San Her dissertation topic was charac- ematics, a program of the Public and Francisco Unified School District terizing signaling pathways triggered Government Affairs Division of the (SFUSD). She also volunteered in neurons after traumatic brain National Research Council. on weekends as a docent at the injury. Prior to attending Penn, Mahlet enjoys spending time with California Academy of Sciences, she had earned a B.S.E. in chemical friends and family, trying out new with a focus on the Philippine coral engineering and M.S.E. in biomedi- restaurants, and traveling. She also reef and American swamp exhib- cal engineering from the University plans to run her first half-marathon its. Through her experiences as of Michigan. In her 10 years of aca- in 2012. SPRING 2012 65

The Asad, Taj, and Jamal Madni Matching Gift Challenges

a 50 percent participation rate by NAE membership. Dr. Madni expressed his hopes for NAE in the following moving words: I am inspired by the brilliant minds at the National Academy of Engineering who provide their expertise to globally improve the quality of life and in addressing issues that are vital to our nation’s future in order to maintain our global technical and economic leadership. These challenges are intended to encourage contribu- tions to the NAE Independent Jamal Madni, Taj Madni, Charles M. Vest, and Asad M. Madni Fund in order to assure the inde- pendence of our voice on national This year, Dr. Asad M. Madni, a active contributors. The Madnis’ policy and to initiate projects that member of the class of 2011, his wife challenge will match, in full, any are vital to our nation’s future. Taj, and his son Jamal are generously gift made in 2012 by members of sponsoring two $50,000 matching the class of 2009, 2010, 2011, or — Asad M. Madni, gift challenges to encourage giving 2012. The Asad, Taj, and Jamal NAE Class of 2011 by NAE members elected from 2009 Madni Section 7 Challenge is focused to 2012 and by fellow members of on increasing the proportion of For more information on participat- Section 7. Dr. Madni approached contributions by members of Sec- ing in the Asad, Taj, and Jamal Madni NAE about sponsoring a challenge tion 7. This challenge grant will Challenges, or to learn more about during the 2011 Annual Meeting. match any individual’s gift that is NAE’s goals for its 50th anniversary, The goal of The Asad, Taj, and larger than that member’s 2011 please contact Radka Nebesky, NAE Jamal Madni Challenge for Newer donation. Both challenges were senior director of development, at Members is to encourage recently inspired by one of NAE’s goals for 202-224-3417 or [email protected], elected NAE members to become its 50th anniversary—to achieve or visit www.nae.edu/giving. The 66 BRIDGE

A Message from NAE Vice President Maxine L. Savitz

Grand Challenges Scholars Program. • initiating crucial new projects Seven Grand Challenges Summits supported by Leadership Gifts have been held in the United States, • promoting friendly competition organized by Duke, USC, Olin Col- among sections to increase the lege, and others, and the first Global percentage of members who give Grand Challenges Summit will be to NAE held in 2013 in cooperation with the Royal Academy of Engineering Nobody understands better than and the Chinese Academy of Engi- our members the crucial role of engi- neering. The 14 Grand Challenges neering and technological innova- Maxine L. Savitz were selected by an NAE-appointed tion in driving the U.S. economy, committee under the sponsorship national security, and quality of I am happy to report that 2011 of NSF, but continued engagement life. If our efforts are successful, was a successful fundraising year for in the project was made possible we will reach our 50th anniversary NAE. Thanks to the extraordinary by unrestricted donations provided with enough resources in hand to generosity and support of more than by NAE members. help the engineering community 650 members, we raised almost $3.4 address the Grand Challenges fac- million in new cash and pledges, 30 50th Anniversary ing us today, turn complex techno- percent of NAE’s annual budget. In December 2014, NAE will logical questions into opportunities We thank you for your confidence celebrate 50 years since its found- for innovation and improvement, that we will use these contributions ing as a parallel organization of out- and maintain NAE’s position as to continue to serve the engineering standing engineers under the NAS a national voice for engineering. community, students, policy makers, charter. We can look back on past Please visit www.nae.edu/giving for and the public. accomplishments spearheaded by more details. More than one-third of the pri- NAE members, including influential vate money we received in 2011, projects and studies, such as Rising Outstanding Contributions slightly more than $1.2 million, Above the Gathering Storm, America’s All contributions are greatly was donated to the NAE Indepen- Energy Future, and the recent report appreciated, and all of them make dent Fund. This vital, unrestricted on the Deepwater Horizon. But we a difference in the work of NAE. money not only provides core sup- can also look ahead to finding new Before I close, however, I want to port, but also enables us to initi- ways of addressing the changing highlight some of the most remark- ate important new projects that needs of a rapidly changing society. able gifts we received in 2011: lack federal funding, as well as to With an eye to the future of • Ross (’02) and Stephanie Coro- expand the scope and impact of NAE, we have launched a 50th tis established a $100,000 chari- current programs. anniversary fundraising activity: table gift annuity to benefit NAE. Perhaps one of the best examples Celebrating 50 Years of Engineering of the exceptional potential of flex- Leadership and Service. We hope • The estate of Lee L. Davenport ible funds was the development of to engage members who have never (’73) established a bequest of the Grand Challenges for Engineer- given to NAE, as well as to inspire $50,000 to support joint work with ing. This initiative, which began as our loyal donors to consider making IOM to improve the efficiency of an idea to inspire student and pub- additional commitments to help us the U.S. health care system. lic engagement with engineering, reach some very ambitious goals: • Irwin (’82) and Joan Jacobs has since taken on a life of its own. • increasing membership in our sponsored two matching gift Several universities have organized lifetime giving societies—the challenges, their third and fourth curricula around the Grand Chal- Einstein, Heritage, and Golden such challenges in the past three lenges and have established the Bridge societies years. This year’s challenges for SPRING 2012 67

the Class of 2011 and the Classes program that promotes effective, Irving T. Waaland (’91) of 2009 and 2010 helped raise innovative engineering education. Johannes (’76) and Julia (’88) almost $150,000 for the NAE Weertman Gifts made regularly each year Independent Fund. Wm. A. Wulf (’93) to NAE also demonstrate genuine • Lockheed Martin donated commitment to our mission and Be on the lookout in 2012 for $500,000 to support the Global goals. As a longtime donor who opportunities to participate in Grand Challenges Summit; understands that every donation to activities leading up to NAE’s 50th $250,000 to EngineerGirl!; and NAE is a choice to support an orga- anniversary celebration. We will be $100,000 to support the National nization whose work matters greatly, discussing plans for this milestone Engineering Forum. I thank the following members who at regional meetings and through- have contributed to NAE for 15 or out the year and will provide more • Asad (’11), Taj, and Jamal more consecutive years: information in our spring and fall Madni have committed $100,000 appeals. Remember, too, that we to sponsor two matching gift Clarence R. Allen (’76) are always available to discuss your challenges in 2012, one for new Charles A. Amann (’89) ideas for celebrating NAE. members and one for Section 7 John C. Angus (’95) On behalf of the NAE Council, members. Wm. Howard Arnold (’74) our president Chuck Vest, and Harold Brown (’67) • Robin (’07) and Rose McGuire myself, thank you very much for Jack E. Buffington (’96) contributed $100,000 to encour- your contributions in 2011. Our Esther M. Conwell (’80) age Native American students to supporting members, friends, part- Irwin Dorros (’90) pursue careers in engineering. ner corporations, foundations, Daniel J. Fink (’74) government sponsors, and other • Gordon (’76) and Betty Moore Robert C. Forney (’89) organizations make all the differ- donated $100,000 to NAE’s Anita K. Jones (’94) ence in our ability to positively un­restricted Independent Fund. Johanna M.H. Levelt Sengers (’92) impact our world and to continue Thomas S. Maddock (’93) • Northrop Grumman donated advocating for engineering. I am Jack S. Parker (’76) $100,000 to encourage diversity grateful for your contributions, and Simon Ramo (’64) in the science and engineering I look forward to your continued William R. Schowalter (’82) workforce. involvement in 2012. F. Stan Settles (’91) • The O’Donnell Foundation Morris Tanenbaum (’72) invested $500,000 in Frontiers of Hardy W. Trolander (’92) Maxine Savitz Engineering Education (FOEE), a Andrew J. Viterbi (’78) The 68 BRIDGE

National Academy of Engineering 2011 Donor Recognition

INDIVIDUALS Lifetime Giving Societies The National Academy of Engineering gratefully acknowledges the following members and friends who have made generous charitable lifetime contributions. Their collective, private philanthropy enhances the impact of NAE as a national voice for engineering.

Einstein Society In recognition of members and friends who have made lifetime contributions of $100,000 or more to the National Academies as personal gifts or as gifts facilitated by the donor through a donor-advised fund, matching gift program, or family foundation. Names in bold are NAE members.

$1,000,000 or more Bruce and Betty Alberts Harry E. Bovay, Jr.* George P. Mitchell Jillian Sackler Norman R. Augustine Donald L. Bren Gordon and Betty Moore Raymond and Beverly Craig and Barbara Barrett Bernard M. Gordon Peter O’Donnell, Jr. Sackler Jordan* and Rhoda Michael and Sheila Held Robert A.* and Mayari Bernard G.* and Rhoda Baruch William R.* and Pritzker Sarnat Stephen D. Bechtel, Jr. Rosemary B.* Hewlett Richard L.* and Hinda Leonard D. Schaeffer Arnold O.* and Mabel Kenneth A. Jonsson* G.* Rosenthal Sara Lee and Axel Schupf M.* Beckman Fred Kavli Fritz J.* and Delores H.* Russ

$500,000 to $999,999 Anonymous Harvey V. Fineberg and Ralph* and Claire* Shela and Kumar Patel Rose-Marie and Jack R. Mary E. Wilson Landau George Rowe, Jr. Anderson Eugene Garfield William W. Lang Jack W. and Valerie Rowe John and Lise Armstrong William T. Golden* Richard and Ronay William J. Rutter Kenneth E. Behring Joan and Irwin Mark Menschel John and Janet Swanson James McConnell Clark Jacobs Ruben F.* and Donna Roy and Diana Vagelos Richard Evans* Thomas V. Jones Mettler Alan M. Voorhees* Cindy and Jeong Kim Dane and Mary Louise Miller

$250,000 to $499,999 John Abelson T. H. Geballe Philip and Sima Stephen and Anne Ryan Richard C. and Rita Penny and Bill George Needleman Henry and Susan Samueli Atkinson Jerome H.* and Barbara Ralph S. O’Connor Melvin I. Simon Warren L. Batts N. Grossman Kenneth H. Olsen* Ted Turner C. Gordon Bell William R. Jackson* Ann and Michael Ramage Leslie L. Vadasz Mrs. Elkan R. Blout Richard M. Morrow Simon Ramo Charles M. and Rebecca Russell L. Carson Anne and Walt Robb M. Vest

*Deceased §Giving matched by the Class of 2011 Challenge ◊Giving matched by the Jacobs Challenge for the Classes of 2009 & 2010 SPRING 2012 69

$100,000 to $249,999 Holt Ashley* Robert and Cornelia Frederick A. Klingenstein Edward E. Penhoet Francisco and Hana Ayala Eaton William I. Koch Percy Pollard William F. Ballhaus, Sr. Tobie and Dan Fink Jill H. Kramer Allen E. and Marilyn Thomas D.* and Janice George and Ann Fisher John W. Landis Puckett Hood Barrow Harold K. and Betty A. Gerald and Doris Laubach Carol and David Richards Elwyn and Jennifer Forsen David M. Lederman Henry M. Rowan Berlekamp William L. and Mary Kay Whitney and Betty Dr.* and Mrs. Joseph E. Diane and Norman Friend MacMillan Rowe Bernstein William H. Gates III Robin K. and Rose M. Maxine L. Savitz David G. Bradley Nan and Chuck Geschke McGuire Wendy and Eric Schmidt Sydney Brenner Paul and Judy Gray William W. McGuire Shep* and Carol Ruth George* and Virginia Corbin Gwaltney Burton and DeeDee Shepherd Bugliarello Margaret A. Hamburg and McMurtry Georges C. St. Laurent, Jr. Malin Burnham Peter F. Brown The Honorable* and Mrs. Edward C. Stone Ralph J. and Carol M. William M. Haney III G. William Miller Judy Swanson Cicerone Jane Hirsh Joe and Glenna Moore Charlotte and Morry A. James Clark M. Blakeman Ingle David and Lindsay Tanenbaum W. Dale and Jeanne Robert L. and Anne K. Morgenthaler Peter and Vivian Teets Compton James David Packard* Andrew and Erna Viterbi Roman W. DeSanctis Anita K. Jones Charles* and Doris John C. Whitehead Robert and Florence Trevor O. Jones Pankow Wm. A. Wulf Deutsch Yuet Wai and Alvera Kan Lawrence and Carol Alejandro Zaffaroni Charles W. Duncan, Jr. Leon K.* and Olga* Papay Janet and Jerry Zucker George and Maggie Eads Kirchmayer Jack S. Parker

Golden Bridge Society In recognition of National Academy of Engineering members and friends who have made lifetime contributions of $20,000 to $99,999 to the Academies as personal gifts or as gifts facilitated by the donor through a donor-advised fund, matching gift program, or family foundation.

$50,000 to $99,999 NAE Members Paul Baran* Robert E. Kahn Darla and George E. Robert F. and Lee S. Lewis M. Branscomb Thomas Kailath Mueller Sproull William Cavanaugh III Paul and Julie Kaminski Cynthia J. and Norman A. H. Guyford Stever* Robert A. Charpie* Theodore C. Kennedy Nadel Leo J.* and Joanne J. John M. Cioffi Kent Kresa Franklin M. Orr, Jr. Thomas Lance A. Davis Bonnie Berger and Frank Richard F. Rashid Johannes Weertman Robert W. Gore Thomson Leighton Charles Eli Reed* Julia Weertman James N. Gray Johanna Levelt Sengers George A. Roberts Robert H. and Joan John O. Hallquist Frank W. Luerssen Warren G. Schlinger Wertheim Michael W. Hunkapiller Robert C. Seamans* Robert M. White Adrian Zaccaria

*Deceased §Giving matched by the Class of 2011 Challenge ◊Giving matched by the Jacobs Challenge for the Classes of 2009 & 2010 The 70 BRIDGE

$20,000 to $49,999 NAE Members Clarence R. Allen Donald N. Frey* Norman N. Li Joel S. Spira William F. Allen, Jr. Edgar J. Garbarini* Asad M., Taj, and Jamal Arnold and Constance William A. Anders Richard L. Garwin Madni Stancell Bishnu S. Atal Louis V. Gerstner, Jr. Thomas J. Malone Chauncey Starr* William F. Ballhaus, Jr. Martin E. and Lucinda James F. Mathis Raymond S. Stata William F. Banholzer Glicksman Roger L. McCarthy Richard J. Stegemeier David K. Barton Joseph W. Goodman Kenneth G. McKay* Stanley D. Stookey Roy H. Beaton* William E. Gordon* James K. Mitchell Daniel M. Tellep Erich Bloch Paul E. Gray John L. Moll* Gary and Diane Tooker Barry W. Boehm Delon Hampton Dan and Patsy Mote Raymond Viskanta Harold Brown Wesley L. Harris Dale and Marge* Myers John C. Wall Corbett Caudill Martin C. Hemsworth* John Neerhout, Jr. Daniel I. C. Wang Joseph V. Charyk John L. Hennessy Robert M. and Marilyn Willis H. Ware Paul Citron Robert and Darlene Nerem William L. Wearly* G. Wayne Clough Hermann Ronald Nordgren Albert R. C. and Jeannie Stephen H. Crandall David and Susan Hodges Simon Ostrach Westwood Malcolm R. Currie Charles O. Holliday, Jr. Donald E. Petersen Willis S. White, Jr. Ruth A. David Kenneth F. Holtby Dennis J. Picard Sheila E. Widnall Gerald P. Dinneen Edward E. Hood, Jr. George B. Rathmann John J. Wise Nicholas M. Donofrio Edward G. Jefferson* Eberhardt Rechtin* Edward Woll* E. Linn Draper, Jr. Min H. Kao Ronald L. Rivest Edgar S. Woolard, Jr. Thomas E. Everhart John and Wilma Will H. Rowand* A. Thomas Young Stephen N. Finger Kassakian Brian H. Rowe* Samuel C. Florman James Krebs Jonathan J. Rubinstein Friend Robert C. and Marilyn Lester C. and Joan M. Roland W. Schmitt Kristine L. Bueche Forney Krogh Donald R. Scifres

Heritage Society In recognition of Academies members and friends who have thoughtfully planned gifts today that will pro- vide for the future by including the Academies in their wills or by other deferred gifts. Names in bold are NAE members. Andreas and Juana Clyde J. Behney Morrel H. Cohen Robert A. Derzon* Acrivos Paul Berg Stanley N. Cohen Peter N. Devreotes Gene and Marian Amdahl Franklin H. Blecher Colleen Conway-Welch Paul M. Doty* Betsy Ancker-Johnson Elkan R. Blout* Ross and Stephanie Mildred S. Dresselhaus John C. Angus Francis R. Boyd* Corotis Merlin K. DuVal* John and Lise Armstrong Daniel Branton Ellis and Bettsy Cowling Ernest L.* and Eva Eliel Norman R. Augustine Robert and Lillian Brent Barbara J. Culliton Gerard W. Elverum W. O. Baker* Corale L. Brierley Malcolm R. Currie Emanuel Epstein Jack D. Barchas James A. Brierley Lee L. Davenport* William K. Estes* H. H. Barschall* Rita K. Chow Henry David* Richard Evans* Stanley Baum John A. Clements Ruth M. Davis Scott E. Forbush* Stephen D. Bechtel, Jr. D. Walter Cohen W. Kenneth Davis* Robert C. and Marilyn Forney

*Deceased §Giving matched by the Class of 2011 Challenge ◊Giving matched by the Jacobs Challenge for the Classes of 2009 & 2010 SPRING 2012 71

Paul H. Gilbert Thomas S. Maddock Daniel W. Pettengill* Robert L. Sinsheimer Norman H. Giles* Artur Mager Nora K. Piore* Arnold and Constance Martin E. and Lucinda Jane Menken Frank Press Stancell Glicksman Stanley L. Miller* Simon Ramo H. Eugene Stanley George Gloeckler Francis D. Moore* Allen F. Rhodes* Dale and Audrey Stein Joseph W. Goodman Gordon and Betty Moore Alexander Rich Rosemary A. Stevens Norman Hackerman* Arno G. Motulsky Frederic M. Richards* John and Janet Swanson Chushiro Hayashi* Van C. and Barbara Mow Henry W. Riecken John A. Swets Michael and Sheila Held Guido Munch Emanuel P. Rivers Esther S. Takeuchi Robert E. Horton* Mary O. Mundinger Richard J. and Bonnie B. Paul Talalay Thomas S. Inui Philip and Sima Robbins Klaus D. Timmerhaus* Richard B. Johnston, Jr. Needleman Richard L.* and Hinda Ivan M. Viest Anita K. Jones Oliver E. Nelson* G.* Rosenthal Willis H. Ware Jerome Kagan Norman F. Ness James F. Roth Robert H. and Joan Samuel Karlin* Ronald Nordgren Esther and Lewis Rowland Wertheim John W. Landis Gilbert S. Omenn Sheila A. Ryan John Archibald Wheeler* Norma M. Lang William and Constance Paul R. Schimmel Maw-Kuen Wu William W. Lang Opie Stuart F. Schlossman Wm. A. Wulf Martin G. Larrabee* Bradford W. and Virginia Rudi Schmid* Charles Yanofsky Thomas M. Leps* W. Parkinson Kenneth I. Shine Michael Zubkoff Edithe J. Levit* Zack T. Pate Dorothea P.* and Herbert R. Duncan Luce Mildred O. Pearson* A.* Simon

Annual Giving Societies The National Academy of Engineering gratefully acknowledges the following members and friends who made chari- table contributions to the Academies during 2011. Their collective, private philanthropy enhances the impact of NAE as a national voice for engineering.

The Class of 2011 Matching Gift Challenge matched, dollar for dollar, gifts made to NAE by NAE members or foreign associates elected in 2011. The Jacobs Challenge for the Classes of 2009 & 2010 matched, fifty cents on the dollar, gifts made to NAE by NAE members or foreign associates elected in 2009 and 2010. Participants in each challenge qualified for the various annual giving societies listed below according to the combined total of their contributions and matching amounts.

Catalyst Society In recognition of members and friends who contributed $10,000 or more in collective support for the Academies in 2011. We acknowledge those contributions made as personal gifts or as gifts facilitated by the donor through a donor- advised fund, matching gift program, or family foundation.

NAE Members Stephen D. Bechtel, Jr. Ruth A. David Arthur M. Geoffrion Joan and Irwin Mark Harry E. Bovay, Jr.* Lance A. Davis Paul and Judy Gray Jacobs George* and Virginia Robert and Cornelia John O. Hallquist Thomas Kailath Bugliarello Eaton Charles O. Holliday, Jr. Min H. Kao§ Robert A. Charpie* Stephen N. Finger◊ Michael W. Hunkapiller Theodore C. Kennedy Paul Citron Tobie and Dan Fink Kent Kresa

*Deceased §Giving matched by the Class of 2011 Challenge ◊Giving matched by the Jacobs Challenge for the Classes of 2009 & 2010 The 72 BRIDGE

Bonnie Berger and Frank Gordon and Betty Moore Simon Ramo Charles M. and Rebecca Thomson Leighton Narayana Murthy and Thomas J. Richardson§ M. Vest Asad M., Taj, and Jamal Sudha Murty◊ Ronald L. Rivest Andrew and Erna Viterbi Madni§ Lawrence and Carol Papay Anne and Walt Robb Adrian Zaccaria Roger L. McCarthy Jack S. Parker George A. Roberts Robin K. and Rose M. Allen E. and Marilyn Maxine L. Savitz Friend McGuire Puckett Joel S. Spira Peter O’Donnell, Jr. Peter and Vivian Teets Rosette Society In recognition of members and friends who contributed $5,000 to $9,999 in collective support for the Academies in 2011. We acknowledge those contributions made as personal gifts or as gifts facilitated by the donor through a donor-advised fund, matching gift program, or family foundation.

NAE Members William A. Anders Joseph V. Charyk Paul and Julie Kaminski Venkatesh Narayanamurti Bishnu S. Atal G. Wayne Clough Robert M. Koerner John Neerhout, Jr. Norman R. Augustine Gerald P. Dinneen Fred C. Lee§ Robert B. Ormsby, Jr. C. Gordon Bell Nicholas M. Donofrio Frank W. Luerssen Bernhard O. Palsson Elwyn and Jennifer Thomas E. Everhart J. Ross and Margaret Henry M. Rowan Berlekamp Robert C. and Marilyn Macdonald Henry and Susan Samueli R. Byron Bird Forney John C. Martin John H. Schmertmann Samuel W. Bodman Nan and Chuck Geschke James F. Mathis Raymond S. Stata Barry W. Boehm Arthur L. Goldstein Dan and Patsy Mote John C. Wall◊ Lewis M. Branscomb Martin C. Hemsworth* Cherry A. Murray Edgar S. Woolard, Jr. Selim A. Chacour◊ Robert E. Kahn Cynthia J. and Norman A. A. Thomas Young Nadel Challenge Society In recognition of members and friends who contributed $2,500 to $4,999 in collective support for the Academies in 2011. We acknowledge those contributions made as personal gifts or as gifts facilitated by the donor through a donor-advised fund, matching gift program, or family foundation.

NAE Members Rodney C. Adkins Eduardo D. Glandt Donald A. Norman§ Jerry Sanders Clyde and Jeanette Baker Paul E. Gray Matthew O’Donnell◊ Linda S. Sanford William F. Baker§ Wesley L. Harris Franklin M. Orr, Jr. Ronald V. Schmidt Becky and Tom Bergman◊ John L. Hennessy Claire L. Parkinson◊ Arnold and Constance M. Elizabeth Cannon§ Keith P. Johnston§ William J. Perry Stancell Albert Carnesale§ John and Wilma Pete Petit§ Richard J. Stegemeier Corbett Caudill Kassakian John W. Poduska, Sr. John and Janet Swanson◊ Sunlin Chou Hau L. Lee◊ Chris D. Poland◊ Raymond Viskanta John M. Cioffi Burn-Jeng Lin Robert H. Rediker Johannes Weertman Joseph M. Colucci Ralph D. Masiello§ Kenneth L. Reifsnider Julia R. Weertman Samuel C. Florman Richard A. Meserve Lisa Dhar and John Willis S. White, Jr. Arthur Gelb◊ Richard B. Miles§ Rogers§ Paul H. Gilbert Ronald Nordgren Mendel Rosenblum and Friend Diane Greene◊ Kristine L. Bueche

*Deceased §Giving matched by the Class of 2011 Challenge ◊Giving matched by the Jacobs Challenge for the Classes of 2009 & 2010 SPRING 2012 73

Charter Society In recognition of members and friends who contributed $1,000 to $2,499 in collective support for the Academies in 2011. We acknowledge those contributions made as personal gifts or as gifts facilitated by the donor through a donor-advised fund, matching gift program, or family foundation.

NAE Members Anonymous Uma Chowdhry Alexander F. Giacco Jack E. Little Andreas and Juana Michael J. Cima§ Herbert Gleiter Christopher B. Lofgren◊ Acrivos Philip R. Clark Mary L. Good J. David Lowell William G. Agnew Esther M. Conwell Joseph W. Goodman Lester L. Lyles§ Alfred V. Aho William J. Cook§ Barbara J. Grosz Thomas S. Maddock James F. Albaugh§ Dale R. Corson Hermann K. Gummel Artur Mager Harl P. Aldrich, Jr. Gary L. Cowger Edward E. Hagenlocker John L. Mason Nicolaos G. Alexopoulos Henry Cox James S. Harris, Jr.§ Robert D. Maurer Clarence R. Allen Stephen H. Crandall George N. Hatsopoulos James C. McGroddy John L. Anderson Natalie W. Crawford Siegfried S. Hecker Kishor C. Mehta Minoru S. Araki James Q. Crowe Joachim Heinzl James K. Mitchell Neil A. Armstrong Malcolm R. Currie Chris T. Hendrickson§ Sanjit K. Mitra Wm. Howard Arnold Glen T. Daigger David and Susan Hodges Duncan T. Moore James R. Asay L. Berkley Davis Thom J. Hodgson Albert F. Myers Thomas W. Asmus Carl de Boor Edward E. Hood, Jr. Dale and Marge* Myers David Atlas Pablo G. Debenedetti John R. Howell Albert Narath Nadine Aubry§ Thomas and Bettie Deen J. Stuart Hunter Robert M. and Marilyn Ken Austin Armen Der Kiureghian§ Mary Jane Irwin Nerem Arthur B. Baggeroer Ralph L. Disney Wilhelmina and Stephen Paul D. Nielsen◊ James E. Barger§ L. K. Doraiswamy◊ Jaffe Chrysostomos L. Nikias David K. Barton Irwin Dorros Leah H. Jamieson Amedeo R. Odoni§ Jordan* and Rhoda Elisabeth M. Drake George W. Jeffs Robert M. Oliver Baruch James J. Duderstadt Barry C. Johnson Bradford W. and Virginia Arden L. Bement, Jr. Susan T. Dumais§ Michael R. Johnson◊ W. Parkinson Daniel Berg Lloyd A. Duscha G. Frank Joklik Shela and Kumar Patel Rudolph Bonaparte Lewis S. Edelheit Anita K. Jones Donald E. Petersen Seth Bonder* Gerard W. Elverum Trevor O. Jones Kurt E. Petersen H. Kent Bowen Thomas V. Falkie Eric W. Kaler◊ Percy A. Pierre◊ Corale L. Brierley James A. Fay Melvin F. Kanninen William P. Pierskalla James A. Brierley Anthony E. Fiorato Sung Wan Kim William F. Powers Rodney A. Brooks Robert E. Fischell James L. Kirtley William R. Pulleyblank◊ Alan C. Brown George and Ann Fisher Albert S. Kobayashi Henry H. Rachford, Jr. Andrew Brown, Jr. Edith M. Flanigen Jindrich Kopecek§ Prabhakar Raghavan Harold Brown Peter T. Flawn Demetrious Koutsoftas Doraiswami Ramkrishna◊ Ned H. Burns Harold K. and Betty A. Lester C. and Joan M. Richard F. Rashid Jeffrey P. Buzen Forsen Krogh Buddy D. Ratner Federico Capasso John S. Foster, Jr. Charles C. Ladd Richard J. and Bonnie B. Francois J. Castaing Elsa M. Garmire Fred J. Leonberger Robbins Jean-Lou A. Chameau◊ Donald P. Gaver◊ Dennis P. Lettenmaier◊ Bernard I. Robertson Chau-Chyun Chen Louis V. Gerstner, Jr. Frances S. and George T. Alton D. Romig, Jr. Ligler Anatol Roshko

*Deceased §Giving matched by the Class of 2011 Challenge ◊Giving matched by the Jacobs Challenge for the Classes of 2009 & 2010 The 74 BRIDGE

William B. Russel Gregory Stephanopoulos A. Galip Ulsoy Wm. A. Wulf Allen S. Russell Henry E. Stone John M. Undrill§ Israel J. Wygnanski Vinod K. Sahney William D. Strecker Harold J. Vinegar Beverly and Loring Wyllie Joseph C. Salamone§ Charlotte and Morry Thomas H. Vonder Haar William W-G. Yeh Harvey W. Schadler Tanenbaum Rob and Robyn Wagoner Paul G. Yock◊ Richard Scherrer◊ George Tchobanoglous Darsh T. Wasan William D. Young William R. Schowalter Matthew V. Tirrell Warren M. Washington Mark D. Zoback§ Jane and Jerome Schultz Haldor F. A. Topsoe David A. Whelan Henry G. Schwartz, Jr. James A. Trainham III J. Turner Whitted Friends Robert F. and Lee S. Hardy W. Trolander Ward O. Winer Evelyn S. Jones Sproull Richard H. Truly Herbert H. Woodson Toby Wolf Gunter Stein James E. Turner, Jr. Richard N. Wright

Other Individual Donors In recognition of members and friends who contributed up to $999 in collective support for the Academies in 2011. We acknowledge those contributions made as personal gifts or as gifts facilitated by the donor through a donor- advised fund, matching gift program, or family foundation.

NAE Members Hiroyuki Abe Arthur E. Bergles Don B. Chaffin E. Linn Draper, Jr. H. Norman Abramson Philip A. Bernstein A. Ray Chamberlain Floyd Dunn Linda M. Abriola Madan M. Bhasin Vernon L. Chartier Ira Dyer Hadi Abu-Akeel James R. Biard Andrew R. Chraplyvy David A. Dzombak Arthur P. Adamson David P. Billington Robert P. Clagett Peter S. Eagleson Ronald J. Adrian Jack L. Blumenthal John L. Cleasby Helen T. Edwards Joseph A. Ahearn◊ Alfred Blumstein Robert E. Cohen◊ Manuel Elices Paul A. Allaire F. Peter Boer Seymour B. Cohn Bruce R. Ellingwood Frances E. Allen Mark T. Bohr John P. Connolly◊ Joel S. Engel Charles A. Amann Geoffrey Boothroyd Richard A. Conway John V. Evans John David Anderson, Jr.◊ George H. Born Paul M. Cook Lawrence B. Evans John G. Anderson Lillian C. Borrone Stuart L. Cooper§ S.M. Farouq Ali◊ John C. Angus Rafael L. Bras Richard W. Couch, Jr. Joseph Feinstein Frank F. Aplan Clyde L. Briant◊ Arthur Coury◊ Robert E. Fenton Ali S. Argon Peter R. Bridenbaugh Lawrence B. Curtis Bruce A. Finlayson R. Lyndon Arscott Yvonne C. Brill David E. Daniel Essex E. Finney, Jr. Eric Ash Frederick P. Brooks, Jr. Delbert E. Day Millard and Barbara Jamal J. Azar Norman H. Brooks F. P. de Mello Firebaugh Donald W. Bahr Howard J. Bruschi Raymond F. Decker Merton C. Flemings Ruzena K. Bajcsy Randal E. Bryant Don U. Deere Christodoulos A. Floudas§ Rodica A. Baranescu Jack E. Buffington Anthony J. DeMaria G. David Forney, Jr. Grigory I. Barenblatt Anne and John Cahn Robert H. Dennard Abdel-Aziz A. Fouad Robert W. Bartlett James D. Callen Robert C. DeVries Charles A. Fowler John W. Batchelor Joe C. Campbell Daniel W. Dobberpuhl Eli Fromm Zdenek P. Bazant Sebastien Candel◊ Robert H. Dodds Douglas W. Fuerstenau Leo L. Beranek E. Dean Carlson Albert A. Dorman Shun Chong Fung Toby Berger William Cavanaugh III Earl H. Dowell Elmer and Jennifer Gaden

*Deceased §Giving matched by the Class of 2011 Challenge ◊Giving matched by the Jacobs Challenge for the Classes of 2009 & 2010 SPRING 2012 75

Theodore V. Galambos Sheldon E. Isakoff Paul A. Libby Devaraysamudram R. Zvi Galil Jeremy Isenberg Peter W. Likins Nagaraj Gerald E. Galloway, Jr. Kenji Ishihara◊ Robert A. Lindeman◊ Tsuneo Nakahara Edwin A. Gee Tatsuo Izawa Kuo-Nan Liou David J. Nash Ronald L. Geer Andrew Jackson and Nathan and Barbara Alan Needleman David B. Geselowitz Lillian A. Rankel◊ Liskov Stuart O. Nelson Don P. Giddens Linos J. Jacovides§ Andrew J. Lovinger Joseph H. Newman Maryellen Giger◊ David Japikse Mark and Mary David Okrent Jacqueline Gish§ Paul C. Jennings Lundstrom◊ Elaine S. Oran Richard J. Goldstein James O. Jirsa Verne L. Lynn David H. Pai Steve and Nancy Donald L. Johnson Noel C. MacDonald Athanassios Z. Goldstein Marshall G. Jones Subhash Mahajan Panagiotopoulos Solomon W. Golomb Angel G. Jordan Thomas J. Malone Morton B. Panish David J. Goodman Aravind K. Joshi Robert Malpas Stavros S. Papadopulos◊ Roy W. Gould John W. Kalb James W. Mar Donald R. Paul Thomas E. Graedel Ahsan Kareem◊ William F. Marcuson III Arogyaswami J. Paulraj Gary S. Grest Linda P. Katehi Robert C. Marini H. W. Paxton William Gropp◊ Michael C. Kavanaugh Hans Mark P. Hunter Peckham Ignacio E. Grossmann Leon M. Keer James J. Markowsky Syd S. Peng Karl A. Gschneidner Chaitan Khosla◊ David K. Matlock Alan W. Pense Jerrier A. Haddad Pradeep K. Khosla William C. Maurer Nicholas A. Peppas Carl W. Hall Timothy L. Killeen Walter J. McCarthy, Jr. John H. Perepezko Carol K. Hall Judson and Jeanne King William J. McCroskey Thomas K. Perkins Robert N. Hall Riki Kobayashi William J. McCune, Jr. Julia M. Phillips William J. Hall Leonard J. Koch William McGuire Mark R. Pinto◊ Thomas L. Hampton Bernard L. Koff Ross E. McKinney Franz F. Pischinger Thomas J. Hanratty Max A. Kohler Robert M. McMeeking Karl S. Pister Niels Hansen Bill and Ann Koros Terence P. McNulty Stephen B. Pope◊ John M. Hanson Mark H. Kryder Alan L. McWhorter Michael Prats Henry J. Hatch Thomas F. Kuech◊ Eugene S. Meieran Ronald F. Probstein Adam Heller John M. Kulicki Angelo Miele Charles W. Pryor, Jr. Martin Hellman Stephanie L. Kwolek James J. Mikulski Edwin P. Przybylowicz Robert W. Hellwarth Bruce M. Lake James A. Miller Gintaras V. Reklaitis Joseph M. Hendrie James L. Lammie Warren F. Miller, Jr. Eli Reshotko Arthur H. Heuer Benson J. Lamp Keith K. Millheim James R. Rice George J. Hirasaki David A. Landgrebe Richard K. Moore John R. Rice John P. Hirth William W. Lang Walter E. Morrow, Jr. Bruce E. Rittmann William C. Hittinger Robert C. Lanphier III A. Stephen Morse Jerome G. Rivard David G. Hoag Louis J. Lanzerotti Joel Moses Leslie E. Robertson and Lester A. Hoel Cato T. Laurencin Gerard A. Mourou Sawteen See Stanley H. Horowitz Alan Lawley Richard S. Muller Lloyd M. Robeson Thomas J. R. Hughes Edward D. Lazowska E. Phillip Muntz Stephen M. Robinson Arthur E. Humphrey Margaret A. LeMone Earll M. Murman Arye Rosen Salim M. Ibrahim Johanna Levelt Sengers Haydn H. Murray Howard B. Rosen Izzat M. Idriss Kenneth Levy Thomas M. Murray Ken Rosen Kenichi Iga Salomon Levy Gerald Nadler Arthur H. Rosenfeld◊

*Deceased §Giving matched by the Class of 2011 Challenge ◊Giving matched by the Jacobs Challenge for the Classes of 2009 & 2010 The 76 BRIDGE

Hans T. Rossby Neil G. Siegel Edwin L. Thomas◊ Jasper A. Welch, Jr. William B. Rouse Daniel P. Siewiorek James M. Tien David C. White B. Don and Becky Russell Peter G. Simpkins Neil E. Todreas John A. White, Jr. Andrew P. Sage Kumares C. Sinha John W. Townsend, Jr.* Robert M. White Eugene C. Sakshaug Jack M. Sipress Charles E. Treanor Robert M. White Jean E. Sammet Ernest T. Smerdon Alvin W. Trivelpiece George M. Whitesides Gurmukh S. Sarkaria Donald M. Smyth Howard S. Turner Robert V. Whitman Peter W. Sauer Gurindar S. Sohi◊ Stephen D. Umans◊ Kaspar J. Willam Thorndike Saville, Jr. Soroosh Sorooshian Moshe Y. Vardi Eli Yablonovitch Robert F. Sawyer Alfred Z. Spector Anestis S. Veletsos Yannis C. Yortsos George W. Scherer Francis M. Staszesky Walter G. Vincenti Les Youd Roland W. Schmitt Dean E. Stephan Irv Waaland Laurence R. Young Albert B. Schultz Thomas G. Stephens Wallace R. Wade§ Xiang Zhang◊ Lyle H. Schwartz James C. Stevens§ C. Michael Walton Paul Zia Mischa Schwartz Kenneth E. Stinson John D. Warner Ben T. Zinn Shirley E. Schwartz Kenneth H. Stokoe II John T. Watson Dusan S. Zrnic Terrence J. Sejnowski§ Howard and Valerie Walter J. Weber, Jr. Hratch G. Semerjian Stone◊ Wilford F. Weeks Friends Robert J. Serafin Richard G. Strauch Mark N. Wegman◊ Marguerite Adams F. Stan Settles G. B. Stringfellow Robert J. Weimer Jeanne M. Fox Maurice E. Shank Stanley C. Suboleski Lawrence M. Wein◊ Sharon P. Gross Don W. Shaw Richard M. Swanson◊ Sheldon Weinig Roger and Dolores Kiel Thomas B. Sheridan James M. Symons Paul B. Weisz Joy Szekely Martin B. Sherwin Rodney J. Tabaczynski Irwin Welber Elizabeth W. Toor

Tribute In Honor of Arthur L. Goldstein – Cabot Corporation Fund In Memory of Laurence J. Adams – Marguerite Adams In Memory of Jordan J. Baruch – Rhoda Baruch In Memory of Gerard F. Fox – Jeanne M. Fox In Memory of Howard Jones, Jr. – Evelyn S. Jones In Memory of Herbert L. Toor – Elizabeth W. Toor In Memory of Jack Keil Wolf – Toby Wolf

FOUNDATIONS, CORPORATIONS AND OTHER ORGANIZATIONS Lifetime In recognition of foundations, corporations, and other organizations that have made lifetime contributions of $1 million or more to the National Academy of Engineering. AT&T Corporation E.I. du Pont de Nemours International Business O’Donnell Foundation S. D. Bechtel, Jr. & Company Machines Corporation The Ohio University Foundation Ford Motor Company Lockheed Martin Foundation The Boeing Company General Electric Company Corporation Robert Pritzker Family The Charles Stark Draper General Motors Company McDonnell Douglas Foundation Laboratory The Grainger Foundation Corporation Alfred P. Sloan The Andrew W. Mellon Foundation Foundation

*Deceased §Giving matched by the Class of 2011 Challenge ◊Giving matched by the Jacobs Challenge for the Classes of 2009 & 2010 SPRING 2012 77

Annual In recognition of foundations, corporations, and other organizations that contributed to the National Academy of Engineering in 2011. A-dec, Inc. Employees Charity Margaret and Ross Qualcomm Foundation Advanced Micro Devices, Organization of Macdonald Charitable Matching Gifts Program Inc. Northrop Grumman Fund of Triangle Qualcomm, Inc. Avid Solutions Industrial Forney Family Foundation Community Foundation Leslie E. Robertson Process Control GE Foundation McGroddy Family Associates The Baruch Fund Arthur and Linda Gelb Foundation Inc. Rouse Associates, LLC Stephen Bechtel Fund Charitable Foundation Microsoft Corporation Henry M. Rowan Family S. D. Bechtel, Jr. General Electric Company Microsoft Matching Foundation, Inc. Foundation Geosynthetic Institute Gift Program/Giving Samueli Foundation Bell Family Foundation Gerstner Family Campaign Tawny & Jerry Sanders Elwyn and Jennifer Foundation Gordon and Betty Moore Charitable Foundation Berlekamp Fund The Glenmede Trust Foundation Skidmore, Owings & The Bodman Foundation Company Dale and Marge Myers Merrill LLP Boeing PAC Match Google, Inc. Fund at The San Diego Southwest Research Program Gratis Foundation Foundation Institute Cabot Corporation H.O.R. Heavy Oil National Philanthropic Ray & Maria Stata Family Castaing Family Technologies Trust Charitable Fund of Foundation Indo-US Science and Northrop Grumman November 1983 The Robert A. Charpie Technology Forum Corporation Morris & Charlotte Foundation Joan and Irwin Jacobs O’Donnell Foundation Tanenbaum Family Continuing Fund of the Jewish The Ohio University Foundation Bioengineering Ed, Inc. Community Foundation Foundation University of California, Cummins Business KAO Family Foundation Parametric Technology Berkeley Services W.M. Keck Foundation Corporation Viterbi Family Grant The Charles Stark Draper Lockheed Martin Donald and Jo Anne Fund of the Jewish Laboratory Corporation Petersen Fund Community Foundation E.I. du Pont de Nemours Lutron Foundation Poduska Family Zarem Foundation & Company Foundation Inc.

We have made every effort to list donors accurately and according to their wishes. If we have made an error, please accept our apologies and contact the Development Office at 202.334.2431 so we can correct our records.

*Deceased §Giving matched by the Class of 2011 Challenge ◊Giving matched by the Jacobs Challenge for the Classes of 2009 & 2010 The 78 BRIDGE

Calendar of Meetings and Events

March 1–31 Election of NAE officers and April 18 NAE Regional Meeting May 10–11 NAE Council meeting councillors University of Maryland May 16 NAE Regional Meeting March 22 NAE Regional Meeting College Park, Maryland Cornell University University of California April 25–26 Meeting of the Committee on Ithaca, New York Santa Barbara, California Integrated STEM Education May 23 NAE Regional Meeting March 29–31 German-America Frontiers of April 28–29 USA Science & Engineering Jet Propulsion Laboratory Engineering Symposium Festival Pasadena, California Potsdam, Germany Walter E. Washington Convention Center, Washington, D.C. All meetings are held in National Academies April 2 Deadline for nominations for facilities in Washington, D.C., unless otherwise 2012–2013 NAE awards noted.

In Memoriam

ERSEL A. EVANS, 89, retired ADEL F. SAROFIM, 77, Presi- HAROLD S. MICKLEY, 93, vice president and laboratory dential Professor, Department of retired vice chairman, Stauffer technical director, Westinghouse Chemical Engineering and Insti- Chemical Company, died on Hanford Company, died on Novem- tute for Clean and Secure Energy, December 3, 2011. Dr. Mickley ber 9, 2011. Dr. Evans was elected University of Utah, died on Decem- was elected to NAE in 1978 “for to NAE in 1975 “for leadership ber 4, 2011. Dr. Sarofim was elected research on transpired turbulent and contributions in the develop- to NAE in 2003 “for advancing our boundary layers, and leadership in ment and application of fuels and understanding of the mechanisms the industrial development of oxy- materials engineering for advanced and modeling of processes that con- hydrochlorination processes.” nuclear systems.” trol radiation in and pollutant emis- sions from combustors.”

NAE Annual Meeting, September 30–October 1, 2012

The 2012 NAE Annual Meeting, on September 30–October 1, 2012, will be held at the National Academies building and the Keck Center of the National Academies in Washington, D.C. Orientation for the Class of 2012 will be held on September 29, the day before the meeting. On the evening of the 29th, new members and foreign associates will attend a black tie dinner in their honor hosted by the NAE Council. The induc- tion ceremony will be held at noon on Sunday, September 30. An awards program will follow in the afternoon. Events for Monday morning, October 1, will include the Business Ses- sion for members and foreign associates and the annual NAE Forum. In the afternoon, section meetings will be held at the National Academies Building and the Keck Center. The concluding event that evening will be an optional dinner dance at the JW Marriott. The flyer for the 2012 Annual Meeting and online registration form will be available on the NAE website in mid-June. The flyer will also provide information about the spouses’ program on Monday, October 1. SPRING 2012 79

Publications of Interest

The following reports have been developing critical science capac- to accomplish today, and describes published recently by the National ity that is now lacking; linking new future prospects for the agency. The Academy of Engineering or the knowledge to practical applications; report, which includes several find- National Research Council. Unless and establishing an overall gover- ings, but no recommendations, is a otherwise noted, all publications are nance structure to enable the pro- foundational resource for USACE, for sale (prepaid) from the National gram to move in new directions. Congress, federal agencies, and Academies Press (NAP), 500 Fifth NAE members on the study com- project co-sponsors, among others. Street, N.W., Keck 360, Washing- mittee were Warren M. Washington NAE members on the study com- ton, DC 20001. For more informa- (chair), senior scientist, National mittee were David A. Dzombak tion or to place an order, contact Center for Atmospheric Research, (chair), Walter J. Blenko Sr. Uni- NAP online at http://www.nap.edu and Robert E. Dickinson, professor, versity Professor of Environmen- or by phone at (800) 624-6242. Department of Geological Sciences, tal Engineering, Carnegie Mellon (Note: Prices quoted are subject to University of Texas, Austin. Paper, University; Gregory B. Baecher, change without notice. Online orders $35.00. Glenn L. Martin Institute Professor receive a 10 percent discount when you of Engineering, University of Mary- sign up for a MyNAP account. Please National Water Resources Challenges land; and Robert A. Dalrymple, add $5.00 for shipping and handling for Facing the U.S. Army Corps of Engineers. Willard and Lillian Hackerman Pro- the first book and $1.00 for each addi- The U.S. Army Corps of Engineers fessor of Civil Engineering, Johns tional book. Add applicable sales tax or (USACE) is responsible for the Hopkins University. Paper, $27.00. GST if you live in CA, CT, DC, FL, construction, operations, and main- MD, NC, NY, VA, WI, or Canada.) tenance of much of the nation’s Recapturing a Future for Space Explora- water resources infrastructure (e.g., tion: Life and Physical Sciences Research A Review of the U.S. Global Change flood-control levees, multipurpose for a New Era. During its more than Research Program’s Strategic Plan. dams, locks, navigation channels, 50-year history, NASA’s success In 2011, a new National Research port and harbor facilities, and in human space exploration has Council committee was formed beach-protection infrastructure); depended on its ability to address a to advise the U.S. Global Change regulating the dredging and fill- wide range of biomedical, engineer- Research Program (USGCRP), ing of wetlands subject to federal ing, physical science, and related which was established in 1990 to jurisdictions; reducing flood dam- obstacles—an achievement made coordinate the efforts of numerous age and supporting commercial possible by committed support for federal agencies and departments navigation; and, since 1996, restor- research in life and physical sci- working on issues related to climate ing ecosystems. The National ences for human space exploration change. The committee’s first major Research Council Committee on and by its human space exploration task was to review the USGCRP U.S. Army Corps of Engineers infrastructures for scientific discov- draft Strategic Plan 2012–2021 on Water Resources Science, Engi- ery. Despite budgetary challenges (“the Plan”). This report includes neering, and Planning was formed, and changes in direction, some numerous suggestions for improving at the request of USACE, to pro- progress has been made. However, the Plan, ranging from relatively vide independent advice on stra- research in the life and physical small edits to basic questions about tegic and planning issues. In this sciences has been scaled back for a the scope, goals, and capacity of report, based on initial investiga- decade, leaving NASA poorly posi- the program to meet its goals. The tions and discussions with USACE tioned either to take advantage of key issues include identifying initial leadership, the committee surveys the fully equipped and staffed ISS steps to enlarge the scope of the key challenges, defines the limits laboratory or to pursue research program, as proposed in the Plan; of what USACE can be expected to support further human space The 80 BRIDGE exploration. Nevertheless, the tion sector, and everyone concerned of the importance of international Committee for the Decadal Survey about the environment, the econ­ scientific cooperation in building on Biological and Physical Sci- omy, and energy security. positive governmental relation- ences in Space, which authored NAE member Jerald L. Schnoor, ships and dealing with sensitive and this report, believes that a focused Allen S. Henry Chair in Engineer- urgent problems. science and engineering program ing and co-director, Department NAE member C.D. (Dan) Mote can be successful. The goal of this of Civil and Environmental Engi- Jr., Regents Professor and Glenn report is to propose a strategy and neering, University of Iowa, was a L. Martin Institute Professor of develop a forward-looking research member of the study committee. Engineering, University of Mary- portfolio that will recapture the Paper, $68.00. land, College Park, was a member excitement of human spaceflight of the workshop committee. Paper, and lead to a U.S. space program U.S. and International Perspectives on $34.00. that benefits the nation, excites Global Science Policy and Science Diplo- the public, and returns the United macy: Report of a Workshop. Based Vision and Voyages for Planetary Sci- States to the forefront of space on many areas of common interest ence in the Decade 2013–2022. This exploration for the global good. among peoples from diverse cul- report provides a survey of the cur- NAE member Vijay K. Dhir, tures, advancing science and engi- rent state of knowledge of the solar Distinguished Professor and dean, neering has increasingly become system and recommends flagship Henry Samueli School of Engineer- a global enterprise. The National missions in planetary science for ing and Applied Science, Univer- Academies held a workshop in 2013–2022 that could provide a sity of California, Los Angeles, was Washington, DC, in February 2011, steady stream of discoveries about a member of the study committee. to assess effective ways to meet the solar system and identifies Paper, $85.75. international challenges through research priorities selected through a sound science policy and science rigorous review based on the results Renewable Fuels Standard: Potential diplomacy. To ensure that the work- of studies by five expert panels. The Economic and Environmental Effects shop would have an international authoring committee concludes that of U.S. Biofuel Policy. The United perspective on these issues, partici- NASA’s highest priority major mis- States has a long history of investi- pants included representatives of sion should be the Mars Astrobiology gating biofuels and is on a course to Bangladesh, Brazil, Egypt, Germany, Explorer Cacher (MAX-C), a mis- substantially increase their use. This Jamaica, Kazakhstan, Malaysia, sion to Mars that could not only help report provides an evaluation of the Morocco, Rwanda, South Africa, determine whether the planet ever economic and environmental con- and Syria, as well as two recently supported life, but could also help sequences of increasing biofuels pro- named U.S. science envoys, Rita answer questions about its geologic duction as a result of the Renewable Colwell and Gebisa Ejeta. This and climatic history. Other major Fuels Standard, as amended by EISA report summarizes issues addressed projects should include (1) a mis- (RFS2). The evaluation includes during the workshop, including the sion to Jupiter’s icy moon, Europa, descriptions of biofuels produced characteristics of science, such as and its subsurface ocean, and (2) the in 2010 and those projected to be its common language and methods; Uranus Orbiter and Probe mission produced and consumed by 2022, the open, self-correcting nature of to investigate that planet’s interior reviews of model projections and research; the universality of impor- structure, atmosphere, and composi- other estimates of relative impacts tant questions; and respect for evi- tion. For medium-size missions, the on land prices, and a discussion of dence. Many workshop participants committee recommends that NASA the potential environmental harm noted that advancing global science select two new missions to be includ- and benefits of biofuels production and science diplomacy (what some ed in its New Frontiers Program, and the barriers to achieving the prefer to call “global science coop- which consists of frequent, mid- RFS2 consumption mandate. The eration”) are distinct, but comple- size spacecraft missions to explore recommendations of the study com- mentary activities. Whether it is the solar system. The committee mittee will be important for policy called cooperation or diplomacy, also cautions that, if NASA can- makers, investors, the transporta- there are many historical examples not conduct these flagship projects SPRING 2012 81

within budget, it should focus on resulting report will be of value to addition, because transportation smaller, less expensive missions first. federal agencies with different kinds accounts for most of the petroleum NAE member A. Thomas Young, of missions, facilities, and operating used in the United States, reducing retired executive vice president, procedures. energy consumption in this sec- Lockheed Martin Corporation, was NAE members on the study com- tor can also bring down the costs a member of the study committee. mittee were David J. Nash (chair), of securing world oil supplies. In Paper, $55.00. chairman and CEO, Dave Nash addition, the transportation sector, & Associates International LLC, which produces one-quarter to one- Achieving High-Performance Federal and Arthur H. Rosenfeld, Law- third of U.S. carbon dioxide (CO2) Facilities: Strategies and Approaches rence Berkeley National Laboratory emissions, can make an enormous for Transformational Change: A Work- (Emeritus). Paper, $38.00. difference in helping to stabilize shop Report. Buildings account for atmospheric concentrations of CO2 40 percent of primary energy use Policy Options for Reducing Energy Use and other greenhouse gases by the in the United States, 12 percent of and Greenhouse Gas Emissions from middle of this century. water consumption, and 60 percent U.S. Transportation: Special Report NAE members on the study com- of all non-industrial waste. In addi- 307. This report focuses on how mittee were John M. Samuels Jr., tion, the production and delivery of policies pertaining to (1) cars and president, Revenue Variable Engi- electricity from power plants for use light trucks, (2) medium and heavy neering LLC, and Kathleen C. in buildings account for 40 percent trucks, and (3) commercial air­ Taylor, retired director, Materials of U.S. greenhouse gas emissions. liners could lead to major changes and Processes Laboratory, General The federal government manages in trends in transportation energy Motors Corporation. Free PDF. approximately 429,000 buildings of use and emissions. These three many types with a total square foot- modes of transportation, the largest Global Change and Extreme Hydrol- age of 3.34 billion square feet world- users of energy, account for the vast ogy: Testing Conventional Wisdom. wide, of which about 80 percent is majority of passenger and freight A major challenge for the climate owned space. More than 30 depart- trips. According to the committee and hydrologic science communi- ments and agencies are responsible that produced this report, reduc- ties is determining the real-world for managing these buildings, and ing petroleum use over the next nature of changes in climate and each agency’s portfolio of facilities half-century will require not only hydrology and apparent anomalies is determined by its mission and tougher fuel economy standards, in reconciling their extreme mani- programs. In 2010, the General but also measures that pique the festations. The National Research Services Administration Office of interest of consumers and suppliers Council Committee on Hydrologic Federal High-Performance Green in vehicle fuel economy, alternative Science held a workshop on January Buildings asked the National Acad- fuels, and a more efficient transpor- 5–6, 2010, to examine how climate emies to appoint an ad hoc commit- tation system. Major policy options warming translates into hydro- tee of experts to conduct a public addressed in this report—fuel taxes, logic extremes, such as floods and workshop and prepare a report that vehicle-efficiency standards, fuel droughts. The workshop brought identifies strategies and approaches standards, infrastructure invest- together three groups of experts— for achieving a variety of objectives ments, and coordinated transporta- atmospheric scientists and hydrolo- associated with high-performance, tion and land-use planning—have gists, whose focus is on the scientific green, federal buildings. To avoid the potential to bring about large underpinnings and empirical evi- repeating information that is already reductions in energy use and emis- dence linking climate variability available in other reports, papers, sions over time, although each and hydrologic extremes, and water books, and databases, the commit- of them also presents challenges managers, that is, decision makers tee decided to focus on how current in terms of the scope and timing charged with designing and oper- initiatives and available resources of results. The committee sug- ating water systems that must be could promote sustainability as the gests that combining transporta- resilient in the face of a changing preferred choice at all levels of deci- tion policy options could achieve climate and hydrologic extremes. sion making. The hope is that the large-scale, timely reductions. In This summary of the workshop The 82 BRIDGE proceedings provides (1) an over- small, untracked debris and meteor- ing facilities will be; (3) and if the view of current science in terms of oids. This report provides a review Astronaut Corps’ fleet of train- climate change and extreme hydro- of NASA’s efforts to understand the ing aircraft is a cost-effective way logic events and (2) a discussion MMOD environment, determines to prepare astronauts for NASA’s of the “conventional wisdom” that what NASA is and is not doing to spaceflight program. This report climate change will “accelerate” the mitigate the risks, and recommends presents an assessment of the issues hydrologic cycle and evaporation improvements. Although the report raised by these questions, but does and generate more precipitation, committee identifies many positive not explicitly address the future of because a warmer atmosphere will aspects of NASA’s MMOD programs human spaceflight. be able to retain more water vapor. (e.g., responsible use of resources), it NAE members on the study com- Descriptions of changes in frequency also recommends that the agency mittee were Bonnie J. Dunbar, and severity of extremes, the diffi- develop a formal strategic plan as a Dunbar International LLC, and culties of modeling these changes, basis for prioritizing the allocation of Henry McDonald, Sim Center, and the problems of communicating funds and efforts. Other recommen- University of Tennessee at Chatta- the results to water resources practi- dations include improving long-term nooga. Paper, $37.00. tioners are also included. modeling and measurements, updat- NAE members on the study com- ing debris environmental models Future Science Opportunities in Antarc- mittee were Dennis P. Lettenmaier, regularly, and taking other steps to tica and the Southern Ocean. Antarc- Robert and Irene Sylvester Profes- improve the characterization of the tica and the surrounding Southern sor of Civil and Environmental long-term evolution of the debris Ocean, one of the world’s last fron- Engineering, University of Wash- environment. tiers, covers nearly 14 million square ington, Wilson Ceramic Laborato- NAE members on the study com- kilometers (approximately 1.4 ry; and Daniel P. Loucks, professor, mittee were George J. Gleghorn times the size of the United States). School of Civil and Environmental (vice chair), retired vice president Establishing and maintaining infra- Engineering, Cornell University. and chief engineer, TRW Space structure to ensure sufficient heat, Paper, $27.00. and Technology Group, and Kyle light, transportation, and drink- T. Alfriend, TEES Distinguished ing water, and at the same time Limiting Future Collision Risk to Space- Research Chair and professor of minimize pollution and ensure the craft: An Assessment of NASA’s Mete- aerospace engineering, Texas A&M safety of researchers requires sub- oroid and Orbital Debris Programs. University. Paper, $45.00. stantial resources. The committee The amount of debris orbiting Earth of experts that conducted this study (aka space junk), which has been Preparing for the High Frontier: The suggests actions the United States accumulating for decades, could Role and Training of NASA Astronauts might take to support the success damage or even destroy satellites in the Post-Space Shuttle Era. The of the next generation of Antarc- and spacecraft. Over the past 50 National Aeronautics and Space tic and Southern Ocean scientists. years, various National Aeronautics Administration’s (NASA) retire- The committee highlights two areas and Space Administration (NASA) ment of the Space Shuttle and of research, studies related to global communities have significantly changing involvement in Inter- change and studies related to funda- improved meteoroid and orbital national Space Station operations mental discoveries. In addition, the debris (MMOD) programs. Satel- will necessitate changes in the role committee identifies key questions lites have been redesigned to protect and requirements of NASA’s Astro- that will drive research in the com- critical components from damage naut Corps. At NASA’s request, a ing decades and highlights opportu- by moving them from exterior sur- National Research Council com- nities for sustaining and improving faces to interior structures. Orbits mittee addresses three main ques- U.S. research efforts in the region. are monitored and adjusted to mini- tions in this report: (1) what the NAE member Thomas F. mize the risk of collision. MMOD role and size of the Johnson Space Budinger, professor of the Gradu- shielding on the International Space Center Flight Crew Operations ate School, University of Cali- Station protects critical components Directorate should be; (2) what the fornia, Berkeley, E. O. Lawrence and astronauts from the impacts of requirements of astronaut train- Berkeley National Laboratory, was SPRING 2012 83

a member of the study committee. tial methodologies for assessing the treating, preventing, and diagnos- Paper, $50.00. risk of proliferation in states that ing bacterial and viral diseases. The host nuclear facilities and how these facility will include a biosafety level Sharing the Adventure with the Public methodologies can be used by deci- 4 (BSL-4) containment laboratory —The Value of Excitement: Summa- sion makers. The major subjects of housed in a 192,000 square foot ry of a Workshop. On November the workshop were nonproliferation building, located in Boston’s South 8–10, 2010, the National Research and new technologies, separate pol- End, an environmental justice com- Council Space Studies Board (SSB) icy and technical cultures, the value munity. Although the BSL-4 labo- held a public workshop on how the of proliferation-resistance analysis, ratory will occupy only 13 percent National Aeronautics and Space and the usefulness of social science of the total space, it has been the Administration (NASA) and asso- approaches. This volume includes a source of virtually all community ciated science and exploration summary of workshop presentations concerns about the project. Numer- communities can communicate the and discussions prepared by a desig- ous public meetings have been held exciting prospects of major NASA nated rapporteur. about plans for the facility, and three activities and programs to the pub- NAE member C. Paul Robinson, legal actions challenging the project lic. Over a period of two years, the President Emeritus, Sandia National have been initiated. This volume, workshop planning committee, in Laboratories, chaired the workshop the fifth in a series, provides infor- conjunction with SSB, developed planning committee. Paper, $36.00. mation about additional studies that five “Grand Questions” in space might be needed to assess the risks science and exploration around Continuing Assistance to the National associated with the siting and opera- which the event was organized. Institutes of Health on Preparation of tion of the building. It also includes This volume includes summaries Additional Risk Assessments for the comments and questions on a pen- of the workshop presentations and Boston University NEIDL, Phase 3. In ultimate (90-percent complete) discussions. 2003, the Boston University Medi- draft of the revised risk assessment, NAE member A. Thomas Young, cal Center was awarded a $128 which according to the authoring retired executive vice president, million grant from the National committee, is much improved over Lockheed Martin Corporation, was Institutes of Health to build one of past documents. Nevertheless, the a member of the workshop planning two National Emerging Infectious committee recommends improve- committee. Paper, $49.00. Diseases Laboratories (NEIDLs), ments before the final public ver- maximum-containment laborato- sion is published. Proliferation Risk in Nuclear Fuel Cycles: ries for research on pathogens. The NAE member John F. Ahearne, Workshop Summary. In 2011, the purpose of NEIDLs is to support Executive Director Emeritus, Sigma U.S. Department of Energy asked the National Institute of Allergy Xi, The Scientific Research Soci- the National Academies to con- and Infectious Diseases bio­defense ety, chaired the study committee. vene a public workshop to address research agenda by conducting Free PDF. the capability of current and poten- research on new approaches to

The Periodicals BRIDGE Postage Paid (USPS 551-240)

National Academy of Engineering 2101 Constitution Avenue, N.W. Washington, DC 20418