Journal of Computing and Information Technology - CIT 10, 2002, 2, 115–124 115

Future Directions of -based Control Systems

M. Coccoli and A. Boccalatte University of Genova — DIST, Department of , Computer, and Systems Science

Recent advances in object-oriented, clientserver tech- software system with full network capabilities nologies, and the Internet are supplying the technology specifically designed for embedded real-time

enablers needed to provide a uniform information archi-  tecture that can be used to build software architecture control systems Open Control System . Such allowing the inter-operation and integration of a wide set a system will exploit the most advanced and of diverse applications. Moreover the emerging standards promising concepts and technologies, aiming start playing a significant role in the shaping of automa- to provide a real-time process control engine tion architectures in enterprises. The inclusion within a

classical control system of Internet-related technologies running on standard platforms both hardware  and open distributed application concepts would give and software commercial off-the-shelf, COTS the present system compliance with current and future thus avoiding the disadvantageous use of pro- technological trends. At the present time, in the field of the industrial automation, real-time embedded control prietary ones. The entire system will strongly

systems more and more need Internet connectivity for leverage on Web related technology, field-bus  operations of remote plant administration, training, and device networks , control networks, and stan- supervisory activities. In this paper the state of the art in dard communicationinterfacing methods, thus embedded control systems is presented within the field of

industrial automation applications, and the technological enabling the system to be used within Com-  scenario is discussed, followed by the trend for the ponent Based Solution Assembly CBSA ar- evolution in process control systems. Guidelines for the design of innovative, thus competitive control systems are suggested. A case study is presented, outcome of an EC project in which one of the authors is involved, where a remote maintenance system is realized.

Keywords: networking, real-time, control systems, embedded systems.

1. Introduction

In this paper, the state of the art and future di- rections of novel Internet-based control systems will be presented together with some guide- lines, which should drive software and hard- ware architecture design for the next genera- tion of process control systems. The proposed target will be achieved by describing a set of components which, joined together, can enable the development of a really open and flexible Fig. 1. Control systems’ layers.

This work has been supported by the European Commission Directorate-General Information Society, contract No. IST-2000-28318 NEOXITE. 116 Future Directions of Internet-based Control Systems

Fig. 2. The evolution of process control systems.

chitectures or together with Laboratory Infor- tion of truly advanced Flexible Manufacturing

  mation Management Systems LIMS . Indeed, Systems FMS . Moreover, the inclusion of In- adoption of field-bus technology will provide ternet technology will ensure full compliance

a flexible and open communication layer for of the system with current trends of informa-  the link between the control level and the de- tion technology IT within process industry vices’ one as it is sketched in Fig. 1. Such towards future control systems architectures, a system, devoted to an effective plant con- where devices will more likely become smart trol, will also embed and demonstrate opti- and autonomous, also embedding web servers misation by means of advanced mathematical as well as connection services. tools, thus providing a modular, flexible, cus- tomisable, and self-sufficient control system for The proposed technical and conceptual scenario even complex processes and applications, out- can be graphically summarised as it is repre- lining the main features for the future genera- sented in Fig. 2, in which an interesting trend is

Fig. 3. The area of process control. Future Directions of Internet-based Control Systems 117

presented, within process control systems, that, integration between these layers cannot be so at first glance, could be considered quite “vi- smooth. A variety of communication protocols, sionary”, but, on the other hand, IT researchers proprietary technologies, and non-standard in- believe it is more likely achievable within five terfaces exist and are widely used. For exam- to six years. In Fig. 3, the project context is ple, current systems use very different networks highlighted, that is the automation level which typology for the interconnection of their three will be the main target for the proposed con- main layers. Such networks, usually propri-

cepts. Process industry community is increas- etary, may show very different characteristics, 

ingly looking for easier and seamless integration from both functional protocols and physical  between these layers, but minimising, even if electric signals points of view, therefore im- not trying to completely skip, the development peding the desired smooth integration and in- of custom code. On the other hand, end users are formation flow throughout the above layers. also wishing and expecting modern integrated Today, only the network systems in the upper plant systems embedding complementary, and level can be considered close to the standard,

usually customerprocess customisable, tools because of the increasing use of and for optimisation such as statistical process con- the success of TCPIP. The other two networks,

trol and advanced control algorithms e.g., sim- specifically the device network and the control

ulation tools, on-line parameter tuning.At network, are still based on more or less propri- present, in most cases, such tasks must be per- etary protocols, even though during the past few years several different field-busses have been formed by external systems e.g., mathemati- cal models and observers running on dedicated adopted, some of them internationally standard- ised, some other being “de facto” standards. PCs as it is not possible for them to be executed on PLCs or even on some DCSs. Currently available technology provides great A control system designed and developed ac- potentiality, which is confirmed by the increas- cording to the guidelines outlined within this ing importance of IT within modern and com- paper will be able to perform standard as well petitive enterprise systems. As regards plant as advanced control mechanisms by means of systems, over the past years the focus has moved a main control engine, including both local from the hardware to the software, mainly due to

the achievable system flexibility. Migration of  and remote human-machine interfaces HMIs ,

standard networking capabilities, and web ser- control capabilities from dedicated devices and  vices. The resulting system will definitely rep- systems e.g., PLCs and DCSs towards stan- dard architectures is fully in line with the men- resent a step ahead towards next generation of tioned technological trend. Moreover, increas- process control systems. After this short intro- ing reliability and performance offered by stan- duction, the state of the art for real-time control dard PC platforms are gaining ground. There- systems will be outlined in Sec. 2. Needed fore, what could be considered visionary some innovation will be enumerated in Sec. 3 while years ago, will be more likely the state of the the migration from PLC to standard PC archi- art within three to four years. Convergence of tecture will be described in Sec. 4. Some tech- different functionality towards a single piece nical highlights will be discussed in Sec. 5, and of hardware is, as a concept, increasingly ac- conclusions will follow thereafter. cepted, even if, from a very practical point of view, the era where PLCs and DCSs, or any process control-dedicated architectures, will be 2. State of the Art in Process Control completely replaced with the new ones, is still Systems far to come. In fact, the industrial community is somehow reluctant with respect to discontin- A traditional control system can be split into uous innovation. three main layers: the field, related to the in- Here, an influential theory is worth mentioning.

putoutput devices, the controller itself, which Well known within evolutionary biology com- is in charge of executing the control logic, and munity, it states that evolution does not gradu- a monitoring, programming, and supervision ally occur as small steps over long periods of level as sketched in Fig. 1. As already men- time, but rather as long periods of stability in- tioned, at present, due to many reasons, the terspersed with rapid changes. These periods of 118 Future Directions of Internet-based Control Systems

rapid changes are often the result of severe envi- cost reduction, time-to-market, enterprise-wide

ronmental changes. This theory is called punc- integration. Up to now, the problem has been  tuated equilibrium Harrold, 2000 . Analysing the lack of technology infrastructure needed to the story of process control systems, there is ev- build a uniform and consistent information ar- idence of the punctuated equilibrium behaviour chitecture that could work for both low-level is this area and over the last few years, Internet- process control systems and high level plant au-

related technology when fully and profitably tomation and business systems. In the process exploited has become the cornerstone for the control area, traditional, closed loop regulation evolution towards a definitely new generation of techniques are thoroughly known, well estab- process control systems. Recently, due to eco- lished, and operational. Still, there is a lot of nomical reasons too, no sensible evolution has room for innovation and optimisation of control

been observed within these systems, in the sense algorithms actually most of them are simple that most of the steps ahead have always been P.I.D.controllers tuned by trial-and-error proce-

strictly related to smooth adaptations and slight dures; in particular with novel low cost proces- modifications of systems rather than to com- sors easily allowing large number crunching and petitive evolutionsrevolutions. On the other very fast data processing, and exploiting fast and hand, it is also true that “control is control”, in reliable network , a real-time the sense that behaviour and general strategies control system can be driven by data from the which can be adopted in controlling a valve are sensors, together with the results of optimisa- quite well defined concepts. Therefore the inno- tion algorithms, identification tools, observers vation should be considered embedded in where on non accessible variables, that could be per- the behaviour is “generated” and managed. formed on-line, in order to get an effective adap- Finally, marketing investigations show that from tive real-time control of the plant. Moreover, a commercial point of view, many companies classic theory on adaptive control can be prof- and software houses state their ability to provide itably used on production plants. In this field, competitive PC-based control systems. Some there is a lot of room for innovation regarding other companies claim to have such products, the adopted control algorithms. With the aim of but further investigation shows that most of improving a plant’s performance, it should also them are still based on proprietary solutions. As be considered the possibility of having an es- a general consideration, integration of very dif- timate on some of the non-accessible variables ferent functionalities that are not present within via identification procedures or even through the mentioned products will be an innovation itself. definition of observers. Control strategies based on optimal predictive control can be considered too. Such a technique will take into account 3. Needed Innovation innovative topics as multi-objective functions and will produce multivariable model predictive As pointed out in the previous section, over the control techniques and tools that can be effec- past few years, industrial automation commu- tively used in embedded systems. Within this nity, the end-users in particular, have been very context, a promising and interesting aspect of careful in adopting new technology, both hard- innovation is the possibility of integrating the ware and software, due to the small amount control part with a simulation environment in in terms of performance enhancement that they such a way that tests on different control strate- could get compared to the drawbacks that an gies or huge fine parameter tuning actions can upgrade of the whole production system could be performed off-line, always with the aim to introduce. Indeed, recently the degree of accep- improve the plant performance, while, through tance concerning the most innovative IT criteria the exploitation of suited numeric integration and approaches has been fast growing. This is algorithms, small memory-foot print simula- so, because confidence in emerging techniques tion environments can be obtained to act on-line has increased, and, above all, these new tech- with the plant control. Actually, there are many nologies can provide an immediate and effec- math tools which are widely used in most re- tive answer to most of the pressing demands search laboratories and they are becoming more from the production that a modern and com- and more popular in industries too; the possi-

petitive industrial company feels urgent, e.g., bility of integrating a real-time control system Future Directions of Internet-based Control Systems 119

with the design phase of the control strategy concepts can only be achieved through the re- should appear a very attractive item to the end- definition of what should be the model for an users, because, in this way, one step less would embedded real-time control system in the direc- be necessary between the design phase and the tion of network services, , flexibility, implementation. Moreover, the use of a simu- and scalability. lation environment will allow decision making founded on the simulated effect of control ac- tions, based upon a what-if basis. That is, the 4. Moving Process Control Tasks from simulator will be used not only in the design Dedicated Platforms Towards phase but also in real-time. It is possible the Standard PCs realisation of interfaces with the available to- day’s simulators as well as the realisation of a small, embedded self-contained own simulator It should be clear, at this point, that the central based on fast computing rather then advanced role standard hardware platforms will play in graphical functions. the near future as well as the fact that the pro- On the side of information technology, con- posed migration implies a lot of issues, ranging tinuous improvement of traditional program- from the provision of real-time and determinis- ming techniques or adoption of a platform- tic performances in run-time mode, up to con- independent language should be performed for trol tasks configuration environments and de- the sake of the source code portability. But, bugging tools. Also the hardware part of the specifically, the reader’s attention should be system should meet specific criteria, such as focused on the ongoing innovation of operat- using silicon disks instead of mechanical ones, ing systems dedicated to real-time applications ability to correctly boot and run even in non-

standard configurations e.g. keyboard, mouse  real-time operating systems - RTOS ,onthe

use of graphical user interfaces, and on the net- or graphical cards not present. Moving con- work connectivity. Given embedded comput- trol tasks from PLC to PC leads to the adop- ers’ unique and varied requirements, a graphic tion of field-busses for the connection of the

plant devices to the soft-control engine. Since  user interface GUI solution must be flexible and the integration of a GUI with the control flexibility is one of the most important features part of the plant should be done always keep- characterizing such systems, a strong flexibility ing deterministic real-time characteristics. An- should be guaranteed as far as the usable field- other interesting point is the development of busses are concerned. Integration of features networked systems where, within client-server other than just the control ones, such as local architecture or distributed systems, office appli- human machine interfaces, archiving capabili- cations can share data with any real-time data- ties within standard common repositories, open base running with the control system. In such interfaces to third party applications and legacy a high-end networked real-time embedded con- systems, and standard networking capabilities troller, Internet applications can be foreseen as are the winning factors of these architectures, well as distributed programming techniques use allowing the users to profitably include PC- will be enabled since the possibility of remotely based control within a larger context and since monitoring a plant is very attractive especially modern computers allow number crunching and

thinking of control systems working in hostile very fast data processing, it is more likely fore- environments e.g., gas, oil, emergency condi- seeable that advanced control algorithms e.g.,

tions. Such a system will be also suitable for optimisation, multiphase loop controllers, ac-

the implementation of virtual laboratories and it curate mathematical models could be executed will be able to offer easy connectivity towards on the same machine, therefore eliminating the

Laboratory Information Management Systems need of specific PCs dedicated to the control  LIMS . In this respect, recent agent-based pro- task, that is money saving. Diverse technical so- gramming techniques could be exploited and lutions should be evaluated, such as performing new paradigms for control systems architec- all the tasks on a single CPU or even distribut-

tures could be defined Ivezic et al., 1999, Jen- ing the workload between the two CPUs on the

nings and Woolridge, 1996. A competitive and same bus. The results would strongly depend effective exploitation of these general-interest on the future evolution of CPUs, their reliability

120 Future Directions of Internet-based Control Systems  and their cost reduction trend as projection for sides, choice of the right I Os is one of the key

future commercial exploitation of the system. points in the design of any control system ar- chitectures: such a choice can even be driven On the other hand, recent advances in object- by efficiency criteria. By exploiting and fur-

oriented, clientserver technologies have pro- vided the technology enablers needed to pro- ther enhancing the field-bus technology that can vide a uniform information architecture that can provide the end-users with a very high degree of be used to build software architecture allowing flexibility in selecting the desired IO devices, the inter-operation and integration of a diverse the users will be enabled to easily compose their set of applications. Moreover, the emergence own customized IO system by selecting what of standards in the domains of operating envi- they consider the best suited products e.g., with ronments, networking, GUIs, and repositories the most convenient performance index andor starts playing a significant role in the shaping prices from the wider and wider range today’s of automation architectures in enterprises. The market is able to offer. So the end-user will be inclusion within control systems of standard definitely enabled to tailor its IO system to the Internet-related technologies and, more gener- real needs of the plant, even according to a cost- ally speaking, open distributed application con- based analysis. Another key point is always the cepts, will guarantee compliance of the system choice of the computing part since the adoption with current and future technological trends, i.e. of standard computing architectures can provide standard networking protocols, full adoption of manifest benefits, guaranteeing the best flexibil- multi-tier architectures strategy, web-based in- ity in the selection of the characteristics of the terfaces, etc. system and in the choice of the components’ suppliers for the entire control system. Industry standard protocols such as Asynchrono-

5. Technical Highlights 

us Transfer Mode ATM , Transmission Con-

 

trol ProtocolInternet Protocol TCP IP , and  In this section, the most innovative aspects of the Hypertext Transfer Protocol HTTP have the proposed architecture are presented. Such been enabling the development and deploy- features are related to the development of net- ment of standard networking architectures so work services, flexibility, scalability, openness, that standard network communication protocols a set of new functionality incorporated into the can be adopted for data exchange. Moreover, control system in order to enhance plants ef- emerging distributed technologies for the link- ficiency and control process quality, a business age of diverse software applications can be re- wide access to information, possible integration garded as powerful means towards the develop- with commercial and trading systems, and, last ment of a real-time control system fully inte- but not least, global reduction of costs. Future grated in a wide office automation and business control systems will show their unique ability to process context. According to all of the above

unify plant-level control and other systems with considerations, it is clearly emerging that an  enterprise resources planning ERP or busi- innovative embedded real-time control system ness systems. Such a theme is becoming in- should be independent on specific solutions,and creasingly important to manufacturers around should rely on standard communication proto- the world. So, it should be offered a com- cols; moreover, it should be easily integrated pletely integrated set of applications in a fully with other standard products, and it should be networked environment with a reliable underly- easily customisable to the user specific requests. ing infrastructure that permits easy linkage with Matching the needs of end users, device sup-

other systems, and also has graphical human- pliers, and software developers too can be ob-  machine-interface HMI programs for moni- tained through a specific tool able to guaran- toring and configuration activities. tee interaction with the many different parts of Flexibility mainly resides in the ability of inte- the whole control system through a complete

grating inputoutput devices and in the adoption set of interfaces, the so called abstraction lay- and integration of standard computing architec- ers. Throughout the adoption of abstraction tures, both from hardware and software points layers technique, the core engine of the con- of view. From the economical and functional trol system becomes insensitive with respect to Future Directions of Internet-based Control Systems 121

particular implementations you are faced with. tegration of hardware and software components, Moreover, a strong interaction with the plant is an operating system should support a wide vari- guaranteed through the field-bus, and the con- ety of processors and be easily portable to both nections with other office applications happen custom and standard hardware.

through TCPIP communication layer. Inter- faces should be provided to enable data ex- Within such a new philosophy, advanced func- change between different software applications, tionalities should be located into single con- and they should permit control strategy design troller boxes. The “box” model could be an and implementation via mathematical models off-the-shelf Personal Computer industrialised and tools possibly based on the use of emerg- to meet reliability and environmental specifica- ing graphical programming languages or specif- tions such as incorporating some HMI function- ality and implementing an advanced software ically designed mathematical algorithms e.g., adaptive control, sliding mode control, model DCS. New functionality to be added to the em- bedded control system should mainly cover the predictive control depending on the context the plant is working in. graphical user interface needs and control capa- bilities enhancement in fact one of the key com- The embedded market has recently changed ponents of the flexibility concept is the ability such that the basic operating system is just to build brand differentiation through the user one in a long list of requirements to be con- interface. Due to the decreasing hardware costs, sidered when beginning the design cycle of a new functionality and the Internet phenomenon, new project. The onslaught of cheaper pro- graphics support for an operating system can cessing power coupled with new powerful pro- become a key building for the embedded de- gramming languages and development tools and vice. Interactive and intuitive graphical user in-

environments, has led to very complex, so-  terfaces GUIs enable the user to more easily phisticated software and to a variety of new control any device and access the features previ- products, including information and Internet ap- ously buried in static “manual” interfaces with pliances, and other consumer-oriented embed- buttons and alpha numeric displays. Given em- ded devices. Today, beyond the core operat- bedded computers’ unique and varied require- ing system technology, any developers require ments, a GUI solution should obviously be flex- a complete package of products and services ible. to successfully build their applications within a strict timeframe. In response to these changes, Finally the proposed design philosophy for an real time operating systems too are evolving innovative control system can be resumed in the and from the monolithic entities of embed- following key-points: ded systems’ early days, they have been grow-

Portability of the core engine on different ing towards today’s multileveled hierarchies, in platforms. which the lowest level is a real-time kernel. Re-

liability, connectivity and scalability, together Scalability with the building blocks for flexibility and cus- tomisation such as portability and GUIs support, Fulfilment of OEM market needs

are mandatory requirements for any innovative Flexibility in selecting the field-bus and the

embedded systems. Due to important price IO system sensitivity, functionality requirements, and re- source limitations, an embedded system should Control and supervisory capability also offer the possibility to be the most scal- Integrated functionality able and configurable such that only the ap- plications’ necessary components will be used. Standard networking capabilities In fact, unlike desktop systems, embedded sys-

Openness to other software products avail- tems may want to run with a limited hardware able on the market budget, for low cost solutions, and thus must make do only with what is needed. Operating The architecture could be the one shown in Fig. 4. system portability will give the developers es- Each box of the above can be regarded as an sential flexibility in application design. Since independent component of the embedded real- embedded computers are a complete custom in- time control system. The assembly of the above 122 Future Directions of Internet-based Control Systems

The software platform will also have to provide

a TCPIP stack enabled for UDP protocol with BSD sockets interface. Moreover, for the sake

of generality, the hardware platform will have to  support 8, 16, and 32 bit signed and unsigned types of variables. IEEE754 compliance should be enough for floating-point variables. The control logic engine software architecture should be designed by using abstraction lay- ers. An abstraction layer is a portion of code allowing customising of the control logic en- gine over different hardware and software plat- forms. In this respect, the abstraction layers shall be modified each time a new configura- tion is foreseen, that is the porting activity to

any particular platformarchitecture. It can be foreseen that drivers for most common config- urations and widely diffused field-busses will be available from hardware suppliers or system Fig. 4. The proposed control system architecture. integrators. Typically, hardware, network and kernel abstraction layers are necessary but it is up to the control logic engine designer the real- components gives raise to a complex and com- isation of a different project which could even plete real-time control system whose kernel is be more articulated and complex depending on integrated by configuration and network facil- the targeted application. ities, graphical programming environments for Based on network technologies, it appears very easy systems’ configuration and development, important to develop a standard communication in addition to advanced tools devoted to process layer in order to allow the embedded controller control and simulation for the definition of the to interact with any other application at differ- best suited control strategies. ent levels of business automation. In order to Only if based on standard components and ar- connect the control system with the rest of the chitectures, such an embedded real-time control world, a set of connectivity services will have system will be able to offer the needed degree to be considered. of granularity, in order to fulfil the most diverse implementations. Development of the Control Logic Engine will have to be oriented to pro- 5.1. Graphical Applications duce not just another soft-logic software but an effective execution engine for a DCS-oriented With the aim of supplying the embedded con- system. By the full integration with all the de- trol system with a graphical user interface, in scribed modules, users will be able to develop a clientserver-based networked architecture, and implement any customised control strate- there should be no constraints on the choice of gies and easily interact with them in a profitable clients typology, in particular for remote moni- way. toring applications, where the portability to dif- From the side of the control logic engine de- ferent platforms has to be considered mandatory velopment, it is obvious that real-time charac- requirement for a system to be open. In order teristics are needed: so, the software develop- to have full portability and compatibility in an ment platform will provide a real-time multi- Internet environment, modern object-oriented tasking kernel which will have to support basic platform independent languages such as Java actions for real-time programming, such as cre- or JavaScript combined with W3C standards ate, delete, suspend, resume on typical real-time HTML, XML and XSL, and cgi-bin or asp tech- multitasking objects like , tasks or nology seem to be more adequate, and should threads, tickers, shared memory, mutex, events. be provided. For example, for the development Future Directions of Internet-based Control Systems 123 of a control panel, it is important to point out pending on the tool used and its characteristics that a web-based application could run on any in terms of accuracy of mathematical models. computers only having a www browser with- By the design of suited numeric integration al- out any additional costs for software licenses or gorithms, a small memory-foot print simulation any particular royalty agreements with software environment can be obtained. Otherwise many companies. Developing a web-based control math tools exist and are widely used in most of panel would be a very low cost solution for the the research laboratories, even in the industries. realisation of an advanced graphical user inter- face which could run in parallel and by-pass, or even replace the traditional industrial con- 6. Conclusions trollers’ control panel.

For the development of the programming and In the presented scenario, a strong need for de- configuration environment too, the graphical in- veloping new technologies and tools support- terface will have to cover a fundamental role ing the fast and reliable implementation of real- and the same considerations as before are still time applications, in particular control systems, valid. In fact the configuration environment as has been observed. For consumer applications, well as the control logic development environ- the main benefit from the adoption of the pro- ment aims to give a user-friendly interface for posed design paradigm can be in cost and effort making the control system description and def- reduction for the design and the development inition easy, and for defining and programming of their embedded real-time control systems, the control strategy. In the phase of description which can be built by means of the integra-

of the control system, the platform and the IOs tion between off-the-shelf hardware and some  are chosen possibly among standard ones and suited software components. For professional the needed device drivers’ code will be linked to applications a sensible reduction of the com- the control logic engine that will run on the con- plexity of a real-time control system can be ob- troller. The control logic development should be tained. Industries, end-users, and control sys- done within a user-friendly environment for the tems vendors are more and more perceiving net- programming of the control logic engine func- worked technologies, embedded solutions, and tionality that is the choice of the best control distributed control as the right way to meet the algorithm depending on the present application above mentioned requirements. Moreover tech- and according to the plant necessity too. nologies and performances are rushing, and we can expect that virtual plant’s concept will be effective in the future and this fact can be seen 5.2. Advanced Control Support and as the first step in the direction of the devel- Simulation Tools opment of virtual laboratory-based industrial research areas. Such a heterogeneous collec- Based on both classical and non-standard con- tion of emerging technologies, ranging from trol theory, different algorithms for on-line pa- distributed data handling and distributed com- rameter tuning and system identification will be puting to the underlying multicast network in- developed by means of standard tools and with frastructure, has the potential to make it possi- the aim to be easily interfaced with the other ble to create distributed scientific and industrial components of the system as well as with other laboratory environments that provide complete third party or commercial tools. By means of location-independent access to instruments and simulations in a suited mathematical environ- analysis resources, thus enabling remote human ment, it will be possible to predict the behaviour collaboration. This possibility can give the abil- of the system in response to particular inputs or ity of experimenting with pilot plants by remote any changes in its control parameters. In order control from different laboratories through the to keep the system open to any possible solu- Internet in such a way that both scientific com- tions, it will be provided with an interface to munity and industry will benefit from informa- tion infrastructures. external simulators and mathematical tools so that the control system designer will be able to The aim of this paper has been the description choose the degree of depth of his analysis de- of future directions and the definition of guide-

124 Future Directions of Internet-based Control Systems  lines for the development of a complete set of ANTONIO BOCCALATTE graduated in Electronic Engineering 1976 at

the University of Genova. In 1987 won a chair at the University of Gen-  software modules components oriented to em- ova, and presently is Professor of “Data Base Management Systems” bedded applications development and deploy- at the Faculty of Engineering. He is involved in research activity on System Architecture and Artificial Intelligence. ment, specifically targeting OEMs embedded

control. In few words the proposed system is

going to promote the convergence of informa- MAURO COCCOLI received the Ph.D. degree in Electronic and Com- tion processing, networking technologies and puter Science Engineering from the University of Genova, 2000. His research activity covers complex non-linear uncertain systems modeling infrastructures. Based on current strength in and control, discontinuous control techniques, simulations activity of access network technologies and embedded sys- industrial plants, software agents, information technology in industrial tems, it can bring together a wide range of essen- automation and web-related activity. tial technologies development related to a really open platform for control systems, that is one of the more interesting topics in this research area. The adoption from the developers’ side of a component-based approach will promote software re-usability. The use of off-the-shelf components will also be strongly considered in the definition of the system architecture which will be based on novel standard hardware com- ponents.

References   1 D. HARROLD, Here we go again — Where con-

trol systems have been, and where they are going, 

Control Engineering, January 2000 .   2 N. IVEZIC, T.E. POTOK,L.POUCHARD, Multi-Agent

Framework for Lean Manufacturing, IEEE Internet



Computing, Vol. 3, No. 5 Sept Oct 1999 . 

 3 N. R. JENNINGS AND M. WOOLDRIDGE, Software  Agents, IEE Review, January 1996 .

Received: November, 2000 Accepted: February, 2002

Contact address: Dr. M. Coccoli University of Genova — DIST Department of Communication, Computer, and Systems Science via Opera Pia, 13 16145 GENOVA, Italy

Phone: 39.010353.2284

Fax: 39.010353.2154

e-mail: coccolidistunigeit Professor A. Boccalatte University of Genova — DIST Department of Communication, Computer, and Systems Science via Opera Pia, 13 16145 GENOVA, Italy

Phone: 39.010353.2284

Fax: 39.010353.2154

e-mail: ninodistunigeit