Control: a Perspective✩
Total Page:16
File Type:pdf, Size:1020Kb
Automatica 50 (2014) 3–43 Contents lists available at ScienceDirect Automatica journal homepage: www.elsevier.com/locate/automatica Survey Paper Control: A perspectiveI Karl J. Åström a,1, P.R. Kumar b a Department of Automatic Control, Lund University, Lund, Sweden b Department of Electrical & Computer Engineering, Texas A&M University, College Station, USA article info a b s t r a c t Article history: Feedback is an ancient idea, but feedback control is a young field. Nature long ago discovered feedback Received 25 July 2013 since it is essential for homeostasis and life. It was the key for harnessing power in the industrial revolution Received in revised form and is today found everywhere around us. Its development as a field involved contributions from 10 October 2013 engineers, mathematicians, economists and physicists. It is the first systems discipline; it represented a Accepted 17 October 2013 paradigm shift because it cut across the traditional engineering disciplines of aeronautical, chemical, civil, Available online 28 December 2013 electrical and mechanical engineering, as well as economics and operations research. The scope of control makes it the quintessential multidisciplinary field. Its complex story of evolution is fascinating, and a Keywords: Feedback perspective on its growth is presented in this paper. The interplay of industry, applications, technology, Control theory and research is discussed. Computing ' 2013 Elsevier Ltd. All rights reserved. Communication Theory Applications 1. Introduction In the modern post-genomic era, a key goal of researchers in systems biology is to understand how to disrupt the feedback Nature discovered feedback long ago. It created feedback of harmful biological pathways that cause disease. Theory and mechanisms and exploited them at all levels, that are central to applications of control are growing rapidly in all areas. homeostasis and life. As a technology, control dates back at least The evolution of control from an ancient technology to a mod- two millennia. There are many examples of control from ancient ern field is a fascinating microcosm of the growth of the modern times (Mayr, 1969). Ktesibios (285–222 BC) developed a feedback technological society. In addition to being of intrinsic interest, its mechanism to regulate flow to improve the accuracy of water study also provides insight into the nuances of how theories, tech- clocks. In the modern era, James Watts' use of the centrifugal nologies and applications can interact in the development of a dis- governor for the steam engine was fundamental to the industrial cipline. This paper provides a perspective on the development of revolution. Since then, automatic control has emerged as a key control, how it emerged and developed. It is by no means encyclo- enabler for the engineered systems of the 19th and 20th centuries: pedic. To describe the field, we have, somewhat arbitrarily, cho- generation and distribution of electricity, telecommunication, sen the years 1940, 1960 and 2000 as separators of four periods, process control, steering of ships, control of vehicles and airplanes, which are covered in sections with the titles: Tasting the Power operation of production and inventory systems, and regulation of Feedback Control: before 1940, The Field Emerges: 1940–1960, of packet flows in the Internet. It is routinely employed with The Golden Age: 1960–2000, and Systems of the Future: after 2000. individual components like sensors and actuators in large systems. We provide a reflection on the complexity of the interplay of the- Today control is ubiquitous in our homes, cars and infrastructure. ory and applications in a subsequent section. It was only in the mid 20th century that automatic control emerged as a separate, though multidisciplinary, discipline. The In- ternational Federation of Automatic Control (IFAC) was formed in I This paper is partially based on work supported by the Swedish Research 1956 (Kahne, 1996; Luoto, 1978; Oldenburger, 1969), the first IFAC Foundation LCCC Linnaeus Center, the ELLIIT Excellence Center, the NSF under World Congress was held in Moscow in 1960, and the journal Au- the Science and Technology Center Grant CCF-0939370 and Contract Nos. CNS- tomatica appeared in 1962 (Axelby, 1969; Coales, 1969). By 2000 1232602, CNS-1302182, and CPS-1239116, and the AFOSR under contract No.FA IFAC had grown to 66 Technical Committees. As a key enabler of 9550-13-1-0008. The material in this paper was not presented at any conference. This paper was recommended for publication in revised form by Editor John several technological fields, control is quintessentially multidisci- Baillieul. plinary. This is clearly reflected in the diverse organizations, AIAA, E-mail addresses: [email protected] (K.J. Åström), [email protected] (P.R. Kumar). AIChE, ASCE, ASME, IEEE, ISA, SCS and SIAM that are included in the 1 Tel.: +46 46 222 8781; fax: +46 46 138118. American Automatic Control Council (AACC) and IFAC. 0005-1098/$ – see front matter ' 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.automatica.2013.10.012 4 K.J. Åström, P.R. Kumar / Automatica 50 (2014) 3–43 There is yet another sense in which control has been mul- history of control (Bennett, 1979, 1993; Bissell, 2009), on individ- tidisciplinary — in its search for theories and principles, physi- ual researchers (Hughes, 1993), and on organizations and projects cists, engineers, mathematicians, economists, and others have all (Mackenzie, 1990; Mindell, 2002, 2008). There are sessions on the contributed to its development. The physicist Maxwell laid the history of the field at many control conferences. theoretical foundation for governors (Maxwell, 1868). Later, one Paradoxically, in spite of its widespread use, control is not of the first books (James, Nichols,& Phillips, 1947) was writ- very much talked about outside a group of specialists; in fact it ten by a physicist, a mathematician and an engineer. The mathe- is sometimes called the ``hidden technology'' (Åström, 1999). One maticians Richard Bellman (Bellman, 1957b), Solomon Lefschetz reason could be its very success which makes it invisible so that (Grewal& Andrews, 2010), and L. S. Pontryagin (Pontryagin, all the attention is riveted to the end product device. It is also Boltyanskii, Gamkrelidze,& Mischenko, 1962) contributed to the more difficult to talk about ideas like feedback than to talk about early development of modern control theory. Indeed, respect for devices. Another reason is that control scientists have not paid mathematical rigor has been a hallmark of control systems re- enough attention to popular writing; a notable exception is the search, perhaps an inheritance from circuit theory (Bode, 1960; 1952 issue of Scientific American which was devoted to Automatic Guillemin, 1940). Control (Brown& Campbell, 1952; Tustin, 1952). Control theory, like many other branches of engineering sci- By 1940 control was used extensively for electrical systems, ence, has developed in the same pattern as natural sciences. process control, telecommunication and ship steering. Thousands Although there are strong similarities between natural and engi- of governors, controllers for process control, gyro-compasses and neering science, there are however also some fundamental differ- gyro-pilots were manufactured. Controllers were implemented as ences. The goal of natural science is to understand phenomena in special purpose analog devices based on mechanical, hydraulic, nature. A central goal has been to find natural laws, success being pneumatic and electric technologies. Feedback was used exten- rewarded by fame and Nobel prizes. There has been a strong em- sively to obtain robust linear behavior from nonlinear components. phasis on reductionism, requiring isolation of specific phenomena, Electronic analog computing was emerging; it had originally been an extreme case being particle physics. The goal of engineering sci- invented to simulate control systems (Holst, 1982). Communica- ence, on the other hand, is to understand, invent, design and main- tion was driven by the need for centralized control rooms in pro- tain man-made engineered systems. A primary challenge is to find cess control and fire control systems. The benefits derived from the system principles that make it possible to effectively understand power of control were the driving force. and design complex physical systems. Feedback, which is at the Although the principles were very similar in the diverse indus- core of control, is such a principle. While pure reductionism has tries, the commonality of the systems was not widely understood. been tremendously successful in natural science, it has been less A striking illustration is that features like integral and derivative effective in engineering science because interactions are essential action were reinvented and patented many times in different ap- for engineered systems. plication fields. The theoretical bases were linearized models and Many overviews of control have been presented in connection the Routh–Hurwitz stability criterion. A few textbooks were avail- with various anniversaries. IFAC held a workshop in Heidelberg able (Joukowski, 1909; Tolle, 1905). Research and development in September 2006 to celebrate its 50th anniversary (IFAC, 2006). were primarily conducted in industry. Automatica celebrates its 50th anniversary in 2014 (Coales, 1969). Control was an established field by 1960 because of its develop- A comprehensive overview of sensors and industrial