Click for next page ( 60


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 59
III The Centers as a Reality Plans, Mechanisms, and Interactions

OCR for page 59

OCR for page 59
Systems Research Center JOHN S. BARAS INTRODUCTION The University of Maryland and Harvard University are very pleased to have been selected for an Engineering Research Center award by the National Science Foundation. On the basis of this award a Systems Re- search Center (SRC) will be established at the College Park campus of the University of Maryland. The focal University of Maryland organi- zational unit participating in the activities of the SRC will be the College of Engineering. Broad participation by several departments is planned: the Electrical, Chemical, Mechanical, and Aerospace Engineering de- partments within the College of Engineering; and the Computer Science and Mathematics departments, along with the Institute for Physical Science and Technology and the Center for Automation Research. The focal Har- vard University organizational unit will be the Decision and Control pro- gram of the Division of Applied Sciences. In this paper I will summarize the research theme and the educational and research programs of the Systems Research Center. In addition, I will describe the planned industrial collaboration program, international program, information dissemination plans, and other aspects of the center. The Research Theme and Its Significance The theme of research conducted at the SRC is to promote basic research in the implications and applications of the three types of technology (VLSI, CAE, and AI)* involved in the engineering design of high-performance, *VLSI = very large scale integrated circuits CAE = computer-aided engineering AI = artificial intelligence 6

OCR for page 59
62 PLANS AND PROGRAMS OF THE EXISTING CENTERS complex, automatic control, and communication systems. Recent ad- vances in computer science (artificial intelligence, expert systems, sym- bolic computation), in microelectronics (VLSI circuits development, availability of computer-aided design tools for special-purpose designs), and in computer-aided engineering (enhanced interactive graphics, pow- erful work stations, distributed operating systems, and data bases) have created a unique environment for innovative research and development in the discipline known as systems engineering. For the purposes of the present paper, systems engineering is defined as the discipline that com- bines automatic control systems and communication and signal processing systems with certain areas of computer engineering. The major research thrust of the discipline at present is the design and implementation of high-performance electronic systems for automatic control and commu- nication. It is appropriate to describe some of the motivational and historical background that influenced our thinking and planning for the SRC. To begin with, the complexity of such systems has recently increased dra- matically. This is manifested, for example, in tighter engineering speci- fications, in the need for adaptation, in requirements for multisensor integration, in the need to account for contingencies (multiple modalities), in totally digital implementations, and in the need for a mix of numerical and logical computations. Some of the challenging design problems that we plan to address in the SRC further illustrate this point: 1. How do we control systems characterized by complex, often poorly defined models? Examples from our program include chemical process control, where often it is difficult to design "correct" loops and equations. 2. How should one automate the operation of systems defined by pre- cise, highly complex simulation models? Problems in flexible manufac- turing systems in our program represent generic examples, wherein time- precedence constraints and the need for adaptive automation further com- plicate design. 3. How should we design systems controlled by asynchronously op- erating, distributed, communicating controllers? Examples from our pro- gram include the computer-aided design (CAD) of computer/communication networks, dynamic capacity allocation in communication satellites, and efficient management of mixed traffic (voice, video, data). 4. How can we develop design tools for real-time, high-performance, non-Gaussian signal processors? Examples from radar, sonar, image, and speech signal processing are found in our program. 5. How can one integrate multiple sensors for robust, digital, feedback control of nonlinear systems? Our program includes many-degrees-of- freedom robotic manipulators with vision, force, and pressure sensors, as

OCR for page 59
PL&AlVS AND PROGRAMS OF THE EXISTING CENTERS 63 well as advanced aircraft flight controllers especially designed for the new generation of unstable aircraft. The SRC will focus on the development of powerful and sophisticated software systems that will help and guide engineers in the design of automation and information-processing systems. The significance of a well-coordinated long-range research program in this critical high tech- nology area is highlighted by the following considerations. First, within the last year the growing role of automation in manufac- turing (flexible manufacturing systems, automated factories, robotics, etc.) has attracted a great deal of publicity as the key to the health of the United States' economy and industry. Second, an information explosion has encompassed the widespread use of computing and communication equipment (including office automation, personal computers, mobile telephone networks, distributed computing systems, sophisticated telephone networks, satellite communications, video discs, video processors, fiber optics channels, and optical storage). Among the scientific-educational community this explosion has reached across the board, from high school to university to research laboratory. More significantly, it has also been extended to the broader public. Third, there is an increasing reliance on automatic control systems to perform precise and demanding tasks in such areas as air traffic control; advanced guidance and control systems (high-performance forward swept- wing aircraft, large space structures, and advanced space satellites); im- proved performance and reliability of power plants; improved control and operation of power distribution systems; sophisticated control devices for computer/communication networks; advanced electronic controllers for robot manipulators and computer vision systems; intelligent autonomous weapons and distributed sensor networks; and distributed decision systems for tactical/strategic management. Unfortunately, currently available theories and design methodologies for such problems are not in synchrony with the currently available or planned implementation media, be it special-purpose chips or computers with specialized architectures and capabilities. More precisely, the avail- able design theories and performance evaluation methods were developed for different (now often obsolete) implementation media such as analog circuits and sequential machines. Although for some problems-admit- tedly a small class it is feasible to develop improved designs using the new hardware capabilities and existing theory, in the majority of problems there is a substantial lag between the available hardware potential and its realization in the systems being built. That gap is precisely where the Systems Research Center intends to focus.

OCR for page 59
64 PLANS AND PROGRAMS OF THE EXISTING CENTERS Of course, there are examples of successful hardware solutions to some of the design problems already mentioned By this I mean the process whereby one adds hardware components, or "boxes," in an ad hoc fash- ion, then tests each addition and adds more components until a satisfactory system is built. I do not believe that a serious argument can be made that this method is a superior one for exploiting the hardware potential available today. On the other hand, substantial theoretical results and knowledge exist in the form of automatic control and communication systems theories that have not been directly linked to hardware implementations. The re- alization that a window of opportunity exists was a major motivating force in planning for the SRC namely, that advances in CAE, VLSI, and AI have made possible the transformation of "paper algorithms" from pow- erful theories into real-time electronic "smart" boxes (Bares, 1981~. A careful reexamination and development of new design theories that in- corporate component hardware advances and the related implementation constraints is long overdue. We can no longer separate the design of a system from the implementation problem. This is a major thrust of the SRC program. The significance of the SRC program can also be illustrated from a financial point of view. Huge investments have been made and will con- tinue to be made for research and development in microelectronics and computer hardware. It is important and prudent to make the comparatively small investment required for the development of design methodologies and software tools that will be used to build systems with this hardware. It is obvious that the sophistication and capabilities of the circuits and devices that we build will be limited by the power of the CAD tools that we use. Thus, the SRC theme encompasses two fundamental components of high-technology industries: automation and communications. It is impor- tant to emphasize that high-technology industries involved in automation and communication directly influence the competitiveness and perfor- mance of more traditional industries. Consider, for example, the influence that advances in automation may have on steel mill operation and auto- motive design and production. This consideration was an important factor in the development of our plans for the SRC. Educational Needs The Systems Research Center aims at the establishment of a strong advanced research and educational program in the above areas. Given the broad knowledge and intellectual background required by the SRC research theme, we have assembled an interdisciplinary team of scientists and engineers from the two universities involved. Members of the team include

OCR for page 59
PL^TS AND PROGRAMS OF THE EXISTING CENTERS 65 electrical, mechanical, and chemical engineers as well as mathematicians, numerical analysts, computer scientists, and microelectronics and artificial intelligence experts. At its projected full operational level the SRC research program will involve some 40 faculty, 120 graduate students, and at least 120 undergraduate students. A large number of students will be influenced by the Center's educational programs. We strongly believe that there is a real need, quite critical for the nation, to educate and train engineering students in the mix of disciplines and knowledge represented by the SRC research programs. A similarly critical need exists for retraining practicing engineers, and this need will be incorporated in our plans. THE RESEARCH PROGRAM The research program for the SRC is an expansion and natural extension of research work already under way by members of our interdisciplinary team. The research activities listed below served as the inspiration and provided much of the motivation for the planning and implementation of the ambitious research goals of the Center. They are in a sense the seeds for interaction and further development of the key ideas behind the con- ception of the SRC. The SRC will provide the fertilized ground for de- velopment of the major thrusts emanating from these early works, which are: optimization-based design in chemical process control perturbation analysis and AI modeling in manufacturing systems symbolic computation and VLSI architectures for the design of real- time non-Gaussian detectors design of a VLSI DFT processor vision sensors and feedback in robotic manipulators. The research program implementation selected for the SRC was influ- enced by three factors. First, the areas of strength of the participating faculty; second, the expected impact of SRC research; and third, a strong commitment to a problem-driven interdisciplinary program. We have as a result selected five focus-application areas to help us measure the success of the basic research program, and to help motivate it by applying the design tools to a diverse set of complex, real-world problems. These areas are described below, together with the currently planned thrusts in each. "Intelligent" CAD of Stochastic Systems We shall combine CAE and AI methods for the design of advanced nonlinear signal processors capable of real-time operation. One thrust is toward the development of expert systems that can "reason" mathematically and understand a variety of signal and system models. The other two thrusts address questions of

OCR for page 59
66 PLANS AND PROGRAMS OF THE EXISTING CENTERS distributed computations in stochastic systems and implementation by "optimal" VLSI architectures. In particular, silicon compilation and spe- cial high-level signal manipulation languages will be studied. Chemical Process Control Here we shall investigate how CAE, AI, and optimization techniques can be applied to the design and control of chemical plants. Modeling and simulation questions will be analyzed and the models built, using the CAD process. In addition, we shall attempt to integrate reliability and safety considerations into the design software and work stations. Telecommunications There are two major thrusts here. The first cen- ters around the development of powerful simulation and CAD systems for computer/communication networks (local-area, flow-control, and recon- figurable networks). This will involve interactive graphics, expert systems, and high-level command languages. The second thrust involves image and speech processing problems and their hardware implementation. Nu- merical and hardware complexity will be studied, as well as fast digital implementations. Advanced Automation and Information Processing in Manufacturing Systems We shall investigate applications of CAE, AI, and optimization. In particular, an integrated program will be pursued that addresses sched- uling problems, adaptive resource allocation, AI systems in manufactur- ing, data-base integration, flexible manufacturing cells, CAD integration in manufacturing resource planning (MRP), optimization-based design, and advanced interactive simulation. CAD of Intelligent Servomechanisms Two major thrusts are the theory and design of an advanced prototype hand-eye machine, and the design of flight controllers for high-performance aircraft. Both involve the in- tegration of many "smart" sensor data and the control of systems with very complicated dynamics, often requiring the use of symbolic algebra for their derivation. Implementations by special-purpose VLSI processors will be examined. In the area of robotics, the program will address pri- marily feedback control of a mechanical hand with many-degrees-of-free- dom, based on integration of data from several sensors. In the design of flight controllers we will focus on optimization-based design for unstable aircraft. The common thread in all these areas is their emphasis on the devel- opment of advanced CAD tools that combine the specific theory and practice of systems engineering with the three technology drivers: CAD,

OCR for page 59
PLANS AND PROGRAMS OF THE EXISTING CENTERS 67 VLSI, and AI. These advanced design methods provide the intellectual bond in this diversified program. The program cuts across the boundaries of a great many engineering and computer science disciplines.* The program is interdisciplinary, problem-driven, and technique-spe- cific. We believe that the fundamental tools, and methodologies for their design, that will be developed as a result of the SRC research program will have a very broad applicability. Furthermore, it is expected that these generic CAD tools will evolve out of strong interaction among the research activities in the five focus areas. Each area includes systems of high complexity and design problems that cannot be attacked by conventional methods. As research progresses in each area we expect to see a cross- fertilization among the various efforts toward development of CAD tools. At the University of Maryland we have already witnessed that phenomenon in design projects on chemical process control and advanced aircraft. Still significant for the SRC's mission is the interaction between the three technology drivers (CAD, VLSI and AI) on the one hand, and, on the other, the disciplines of control and communication systems as rep- resented in the five focus-application areas. It is anticipated that the broadly interdisciplinary program will prompt a fundamental reexamination of control and communication systems theory and methodology. Further- more, it is expected that this latter interaction will foster a secondary level of interaction among the focus areas as hardware implementations for different applications are analyzed and compared. Thus, the research program of the SRC will have two major components: in-depth investigation of the impact of VLSI, CAE, and AI basic research in modeling, mathematical analysis, optimization, computational and numerical methods, control systems techniques, com- munication system techniques, and computer engineering techniques. The first component will address the following matters. Regarding VLSI (the implementation medium), we shall investigate algorithmic and ar- chitectural aspects of VLSI; signal processing chips; and control chips. The design methodologies to be developed must account for VLSI im- plementation constraints. Regarding CAE (the implementation environ- ment), we shall investigate the effects of interactive graphics, interfaces, *The disciplines include: chemical process modeling, polymers, bioreactors, chemical re- actors, aerodynamics, flight controllers, robotic manipulators, vision, sensor design, signal pro- cessing, communication networks, information theory, coding, optimization, control systems design, stochastic control, detection and estimation, algorithmic complexity, algorithm archi- tecture, VLSI array design, optimization-based CAD, numerical linear algebra, numerical math- ematics, rule-based expert systems, knowledge-based expert systems, computer algebra, stochastic processes, queueing systems, manufacturing, and mechanical machining.

OCR for page 59
68 PLANS AND PROGRAMS OF THE EXISTING CENTERS etc., in the design of sophisticated CAD systems. For example, in de- velopments related to the DELIGHT Marylin system (a powerful optim- ization-based design system we use at Maryland), the fact that advanced graphics were to provide the output enabled the numerical analyst to develop an interactive procedure that could handle multi-objective optim- ization. In addition, this environment permits the engineer to study a design problem in his own language, without being overburdened with compli- cated computer procedures. Regarding AI, we shall investigate the effects of symbolic computation and knowledge-based systems on design. The second component is needed because sophisticated new theories and methodologies are required in order to extract the maximum benefit possible from advances in microelectronics, CAE, and AI. As Roland Schmitt (1984) describes the situation: "In the technology of controls, . . . fundamental theoretical advances are needed to catch up with the speed and power of microelectronics." The impact of VLSI technology on signal processing and automatic control systems is emerging as very influential. However, for success in this direction very advanced CAD tools must be developed and popular- ized. The rapid developments we have seen in VLSI chip design and production were made possible by the development and rapid dissemi- nation of precisely such advanced CAD tools. The SRC program aims at producing similarly sophisticated CAD tools in the general area of control and communication systems engineering design. An important factor in future systems engineering theories and design techniques will be the development of expert systems for CAD (Stefik and de Kleer, 1983~. In applying expert systems to design tasks the idea is to pit knowledge against complexity, using expert knowledge to whittle complexity down to a manageable scale. It is anticipated that expert sys- tems will eventually be applied in many design areas; but their use in digital system design, particularly in CAD, will be a major advance. The planned SRC program will develop a broad research activity in this area. AI and symbolic computation promise to revolutionize design. There are very sound reasons for this prediction. First, the cost of generating special-purpose Fortran-based codes is fast getting out of hand. Massive investments in design tools can become either a brake on innovative designs or an argument against further development. AI symbolic com- putation transfers mathematical models of the physics of the system being designed from the code side (applications code) to the data side of the system, where they can be used, manipulated, shared, modified, and even created by the system as easily as numerical data elements. This transfer is essential for the attainment of cheap, easily reconfigurable design tools. Second, AI and symbolic computation prevent the designer's entrainment in specific design procedures and processes provided by custom-coded

OCR for page 59
PLANS AND PROGRAMS OF THE EXISTING CENTERS 69 Fortran programs, and thus allow for a very flexible approach to the design. Symbolic manipulation has immediate and powerful applications in CAD. For example, the amount of nuisance programming required to develop and maintain large design packages can be reduced to a practically neg- ligible part of the overall code. Furthermore, increasing the level of ab- straction at which data and code are specified reduces significantly the complexity of code transportability. Finally, symbolic manipulation per- mits entire mathematical models- logic as well as its numerical param- eters to be treated as data capable of being manipulated, examined, and modified, as well as being executed like a Fortran subroutine. Further advantages offered by AI include natural language processing, automatic deduction, cognitive models, and learning and inference. An excellent example for an application of AI and symbolic computation in aerospace design is given in Elias (1983~. Systems engineers today are called upon to solve complex control and communication design problems for systems often described by huge sim- ulation models. The traditional approach has been to reduce the complexity to a small number of mathematical equations and eventually apply rather simple elements of available theories. Clearly we can do much better than that if we utilize the full power of techniques from CAD and AI. Fur- thermore, the speed provided by VLSI arrays promises to support the often real-time processing need of advanced control and communication sys- tems. For systems of the complexity seen today it is often difficult to write and manipulate the governing equations correctly. Think, for example, of the task facing a chemical engineer who is trying to describe a complex industrial chemical process, starting from simple, elemental chemical re- action equations. His final goal is to design a process controller. Or consider the aerospace engineer who is developing a mathematical model for a large, complex, multi-body, flexible structure in space. Again, his final goal is to design a controller. Both have to manipulate a large number of equations (often more than 100) of different types (algebraic, differ- ential, partial differential, Boolean, etc.~. Symbolic manipulation and rather elementary AI techniques (such as search heuristics, "sup-inf" decision procedures, etc.) can readily reduce these tasks to routine and permit the engineer to concentrate on the design issues. More generally, there is clearly an established need for utilizing AI methodologies in CAD. In the design of flexible manufacturing systems, for example, one en- counters coordination problems that can benefit enormously from the use of automated reasoning programs. The latter can supervise the lower-level numerical CAD programs. To ask such a systems engineer to solve the complex design problems of today without such a combined arsenal of tools is similar to asking a VLSI chip designer to design the chip without the expert CAD tools now available.

OCR for page 59
126 INFORMATION AND TECHNOLOGY EXCHANGE AMONG THE CENTERS Written transfer (publications). The use of newsletters, project sum- maries, and electronic and conventional mail can be effective. The pos- sibility of developing new journals perhaps on computer disks-on various cross-disciplinary engineering subjects should also be considered. Verbal transfer (seminars/symposia workshops). We do not expect computer networks ever to replace these important face-to-face discus- sions. The networks of exchange probably should not be limited to the estab- lished Engineering Research Centers and those to follow. The ERCs should be closely connected with other institutions. For example, many engi- neering schools do not have extensive research activities, but graduate a large proportion of American engineers; for this reason they are sometimes referred to as predominantly undergraduate institutions. Meeting the over- all goals of the ERCs with respect to national competitiveness will require a favorable working relationship of the Centers with some of these insti- tutions. Numerous methods of involving industry people, in both research and teaching, will be tried. Kettering said, in commenting on which fuel was best for the auto- mobile, "Let the engine decide." In situations involving other institutions and organizations, we should "Let the Centers decide." It is clear that we are going to have to use a variety of mechanisms to extend the benefits of the Engineering Research Centers to engineering schools across Amer- ica. BASIC PRINCIPLES Where does all this leave us? At this juncture it leaves us with more questions than answers. The important thing is that we do not overlook any of the important questions as we move ahead. 1. What are the information and management coordination needs of the Centers? 2. What types of networks and management coordination mechanisms will best meet those needs? 3. Who will use the networks, and who will participate in the man- agement coordinating groups if they are established? 4. What criteria for access will be used, especially for universities and industries that are not participating in the funding of the Centers? 5. What are the best techniques or mechanisms to use in determining potential users of Center research and educational program results? Emphasis should be given to a point that has occurred to me repeatedly as I have considered the matter of information and technology exchange

OCR for page 59
CARL W. HALL 127 among the Centers and between the Centers and their participants. It is not a new idea. Justice Oliver Wendell Holmes put his finger on it a long time ago when he said, "Having science in the attic is fine, so long as you remember to use common sense in the living room." The success of the Centers will depend in large measure on the application of a great deal of common sense in their day-to-day operation. All of us- Center management, the NSF, industrial participants, and others who seek to take advantage of the research and educational potential of the Centers must use common sense unsparingly. For example, a prime purpose of the Engineering Research Centers is to develop fundamental knowledge that will give U.S. industry an edge in the race for better and improved technologies. If NSF were to attempt to establish safeguards over the transfer of information to protect U.S. interests, there could evolve such a snarl of paperwork that the Centers could be rendered ineffective even before they got started. We are counting on all the participants to use the rule of reason so that U.S. interests are served. To the maximum extent possible, NSF is pledged to avoid issuing guidance papers and other such directives that could impede and frustrate the ERCs instead of helping them to achieve their intended purpose. CONCLUSION It would be interesting to contemplate what Charles Kettering might say about the Engineering Research Centers. I know that he would be in favor of university-industry cooperation, as he promoted this practice in his own activities. I know that he would be in favor of cross-disciplinary research, as he received engineering degrees in two different fields. I know that he would favor innovative approaches, as he did when he went against the conventional wisdom in using a small motor to operate the cash register. I know that he would urge people to think-an attitude to be en- couraged by the Centers. Once when asked to what he attributed his success in innovation, he explained it this way: "As a youth, I had trouble with my eyes [in fact, he stayed out of school a while], so I couldn't spend a lot of time reading books and papers which said a thing couldn't be done. " Now I know that he read and studied a lot. What he was really saying was: also THINK and ACT. I know that he would favor involving students in real-life situations. He once said, "It's one thing to produce something in the laboratory test tubes and another to manufacture it by the ton." I know that he would favor using the experimental approach, as he did when he said "Let the engine decide."

OCR for page 59
128 INFORMATION AND TECHNOLOGY EXCHANGE AMONG THE CENTERS I believe that Charles Kettering would be a strong supporter and sales- man for the ERC concept. The story is told that at one point Kettering had a difficult time getting the production people to accept a fast-drying paint which, he knew, would greatly accelerate the manufacturing of automobiles. He took an important vice-president to lunch. "Now," he said to the vice-president, "if you could have another color of car, what would you select?" "Blue," came the answer. And at a signal the painters painted the V.-P. 's car blue. After a quick lunch they returned to the car, and there it was beautiful, blue, and dry. Kettering made his sale. I hope we have "made our sale" of the Engineering Research Center concept. It is an important purchase for the nation to make. DISCUSSION There was some discussion of the possibilities for networking and data exchange with respect to the Centers. Dr. Hall noted that each Center will determine its own networking program, but he would expect each Center to involve relevant sectors of industry in the network. Dr. Pipes com- mented that plans for this are already under way in each ERC; he gave the example of a "dial-up" service at the University of Delaware Center, which, when in place, will make data of all kinds available to participating companies at any time.

OCR for page 59
New Factors in the Relationship Between Engineering Education and Research TERRIER A. HADDAD It is taken as an article of faith that research ensures vitality and com- petence and thereby improves the pedagogical ability of faculty. However, this faith is not shared by everyone. There are those who subscribe to the "Mr. Chips" school of thought. In their minds, teaching ability is some- how separate and independent from the subject at hand, that is to say, "a good teacher can teach anything." In the engineering area this argument is further complicated by the dichotomy between "practitioners" and "academics." More than any other profession, engineering must rely for its continuing renewal on the 2 percent of its number who fundamentally do not practice, except for whatever engineering research they may do. Especially since World War II, faculty members have increasingly held the Ph.D., and have been selected for tenure only if they could show outstanding research capabilities. There is probably no set of issues that can stir more emotion than these at meetings of university trustees. Dis- cussions about the relationship between research and teaching ability or the difference between the academic and the practitioner have all the elements of an intellectual donnybrook. Can we strip away the emotional content of the debate and get to the heart of the matter? Most certainly! To begin with, there is simply too much evidence supporting the notion that an engineer or academic who does good research makes a superior teacher. Are there good teachers who do not do research? Certainly! Are there good researchers who are bad teachers? Certainly! How many good researchers are bad teachers? In this day of faculty evaluation by students and tenure procedures that evaluate teaching ability, there are not many. More often than not, student 129

OCR for page 59
130 RELATIONSHIP BETWEEN ENGINEERING EDUCATION ED RESEARCH evaluations of teaching ability and administrative evaluations of research ability point to the same people. The issue of the academic versus the practitioner is getting more com- plex, however, for a number of reasons: Increasingly, practitioners must rely on the latest scientific knowledge to be competitive. This puts the practitioner in the position either of doing engineering research or of being in close touch with researchers. Most researchers who communicate with a range of industrial practitioners are career academics or governmental employees. Industrial researchers are much more constrained. Engineering technology is progressing at a very fast rate, both in academe and in industry. Thus, to stay well informed engineers in industry must communicate with academics and vice versa. (Getting out of date is not exclusive to industrial practitioners.) Engineering practitioners in government and industry specialize along many dimensions in addition to that of their primary engineering discipline. Their jobs will be in such diverse areas as applied research, product development, manufacturing, manufacturing research, manufacturing en- gineering, field engineering, engineering or manufacturing operations, service or maintenance, or a host of other engineering specialties. Even these jobs differ substantially in technical content depending on the given industry. This complexity of the engineer's job content makes relating to engineers in faculty positions quite difficult. Engineers in government or industry truly live in different worlds from their colleagues on faculties. It should come as no surprise that academe and industry are two very different cultures with different values and vastly different practices. It is a matter of some urgency that both groups learn more about each other, become more knowledgeable regarding each other's problems and depen- dencies, and, especially, learn how best to interact so that each can benefit from the other's empathy as well as its technical contribution. This is really a very important matter. Academics educate our successors and are the primary source of research that fuels the engineering engine. Practitioners do little research, but do most of the engineering work that fuels our economy, keeps us domestically and internationally competitive, and advances our manufacturing. The engineer in practice gets results in the most scientific manner possible. More often than not, however, project success is attained pragmatically, and, being based on insufficient knowl- edge, may contain surprises, sometimes of disastrous proportions. Such surprises point the way for further research, and so engineering leads to research just as research leads to engineering. The problem is to devise means that enable the academic researcher and the industrial practitioner to complement each other best without either

OCR for page 59
JERRIER A. HADDAD 131 having to forsake his own world or invade the other's. Clearly, the En- gineering Research Centers (ERCs) were devised as one solution. How will they affect engineering research and education? THE ERCs' EFFECT ON ACADEMIC SEARCH AND EDUCATION ERCs should greatly influence academic research. Industry's heavy participation should help communicate to researchers the problems of execution that stand in industry's way. While many of these problems would have been communicated to the campus in any event, the ERCs will clearly expedite the process and help ensure that the "two culture" syndrome does not slow or block the transmission. Many industry puzzles have stimulated research programs in academe, yielding beneficial results. To the degree that the ERCs can contribute to this process we will all benefit. Optimistic as we may be about the ERCs, we should not expect them to be a cure-all. To begin with, they can only involve a fraction of the faculty. Those faculty members in fields removed from the focus of an ERC will receive only fleeting benefit from the presence of that ERC on campus. Nonetheless, the values and practices in evidence at the ERC will be communicated through faculty club discussions, luncheon con- versations, and cocktail party chitchat. It is a stated objective of the ERC program to involve both undergraduate and graduate students-in the Centers' work. To the degree that this is done, those students will benefit greatly. This is a form of interning. The Committee on the Education and Utilization of the Engineer (CEUE) has concluded that all engineering students should have some form of interning since it has such a positive effect on the student's attitude toward the university experience (NRC, 1985a). Not only does interning bring in a practice component, but it also makes the students see the value of the knowledge they gain from their studies. Interning nurtures personal char- acteristics that come mainly from experience: positive attitudes, interests, values, needs and motives, and affective skills. These skills are listed as the most important interning goals even over technical knowledge by students, graduates, faculty, and supervisors. It is also an objective of the ERCs that the industry people assigned to them provide a two-way connection to industrial activities, moving campus research results to industry and industrial nonproprietary results to the campus. There is little doubt that ERCs will expedite this two-way com- munication. However, we should not lose sight of the fact that the industry people assigned will come from companies' applied research sections. Generally speaking, these individuals are quite far removed from the marketplace on the one hand and from the production operation on the

OCR for page 59
132 RELATIONSHIP BETWEEN ENGINEERING EDUCATION AND RESEARCH other. In general there has been little difficulty in arranging liaisons be- tween campus and industrial researchers. The problem has been and con- tinues to be putting campus researchers in touch with industrial professionals close to the market or the manufacturing scene. We should not expect the ERCs to have much effect in this regard. SUPPORT OF THE UNDERGRADUATE SCHOOLS There is a much more serious problem, however. It is not a problem unique to the ERC program. Rather, it involves the widespread and laud- able practice of rewarding already excellent institutions with further op- portunities to increase their excellence. It is hard to argue against this practice, and I certainly do not mean to. However, it completely ignores the more than 200 engineering schools that mainly educate undergraduates and that need help perhaps even more than the comparative handful of research institutions. It is a fact that schools that award 14 or fewer Ph.D.s a year award close to half the nation's B.S. engineering degrees (NRC, 1985a). The CEUE recommends that we invent better ways to support the undergraduate programs in this second tier of schools. The two tiers are a relatively new phenomenon, having come about largely after World War II as a result of contracting from the mission agencies and the newly created National Science Foundation. The largely undergraduate schools seem to be quiet institutions, lacking influence in the technical community and in government and industry. Nonetheless, they are important to the nation and add considerably to the diversity and richness of our engineering education system. It is certainly worthwhile to consider creative ways of improving their situation. A large problem in this regard is how to give the proper support without rewarding mediocrity and encouraging com- placency. These schools need help, but we must take care to help in ways that lift the standards and level of education. How to accomplish this is a tough problem that is yet unsolved. There are state programs of support for undergraduate schools that seem to be working rather well. Consideration could be given to having similar national programs. I will mention two New York State programs that differ in that one is focused on the student while the other is focused on the institution. The first is called the Tuition Assistance Plan, or TAP. TAP provides assistance to students based on financial need, if they are New York State residents and attend schools in that state. The great merit of this program is that the schools must be attractive to students. Students have a way of picking the best school within the range of their ability to pay. TAP does not attempt to distinguish the relative quality of the various schools, and

OCR for page 59
JERRIER A. HADDAD 133 also leaves untouched the different costs of state-supported and indepen- dent schools. Thus, students of a wide range of abilities are able to attend a wide range of schools. One of the great features of higher education in this country is the continuum of quality that is available to students. In my view, we must beware of any scheme, no matter how attractive, that stratifies higher education by means of a bureaucracy. I have much greater faith in the workings of a free marketplace that allows students to pick the programs best suited to their individual needs. Schools selected by TAP students are free to use the tuition money to do what they deem will make the particular school more attractive to students. As long as the burden of paying all tuition is not placed on the student, tuition costs can rise closer to the tuition the school actually needs in order to attain the excellence of instruction it seeks in the manner it judges best. With accreditation guaranteeing minimum quality, a diversity of schools can best satisfy the nation's needs. If, as a society, we judge that the accreditation minimum should be raised, the matter can be dis- cussed with the Accreditation Board for Engineering and Technology (ABET), which is composed of public-spirited engineers from a cross section of professional societies. The other assistance plan is called the Bundy Plan. Some years ago, while McGeorge Bundy was president of the Ford Foundation, he was asked by the state of New York to recommend a way to keep alive the colleges in the state. Then as now higher education was having its prob- lems. As implemented, the plan gave "Bundy Money" to degree-granting institutions according to the number and types of degrees they annually granted. In the beginning each bachelor's degree earned $400 for the college, each master's degree $400, and each Ph.D. $2,400. These amounts have been increased from time to time, until for the 1985-1986 academic year they will become $1,500 for the bachelor's degree, $950 for the master's degree, and $4,550 for the Ph.D. In addition, the two-year as- sociate's degree warrants $600. Here again the attraction of the plan is that the schools must use the funds to continue attracting the students they need for the degrees they want to grant. If standards are lowered to maximize the number of degrees granted, then the most talented students will stay away. If they are lowered significantly, then ABET will withhold accreditation. A powerful moti- vator is the attractiveness of graduates to graduate schools or the job market. In a free economy you cannot fool the marketplace for long. A significantly different approach would be to have programs aimed at giving each accredited engineering college at least some research funding. Proposals could be judged by people having no connection with the pro- posing college. A minimum amount, perhaps based on the student pop- ulation or faculty size, could be given to support the school's most deserving

OCR for page 59
134 RELATIONSHIP BETWEEN ENGINEERING EDUCATION AND RESEARCH faculty for their competitive standing at that school, independent of na- tional competition. In doing this we would be encouraging the most tal- ented department to raise its standards, thus fulfilling the goal of improving the preparation of engineers. Further, other faculty members would un- doubtedly be stimulated by the competition and seek to improve their research proposals and programs. I know that this flies in the face of the peer review system, which officially ignores the institution. However, our aim is to improve the preparation of our engineers, and to do this we must improve the institutions in the second tier, those educating half of the nation's engineers. If research will improve the pedagogical skill of the faculty, that purpose is just as valid and important to us as the more accepted purpose of adding to fundamental knowledge. A National Science Foundation program comes quite close to what I have just described. The program, called Research in Undergraduate In- stitutions, is only a year or two old, and it seems to be successful in many regards. It is designed to give awards to the smaller schools that are predominately undergraduate. "Predominately undergraduate" is in this case defined as granting 20 or fewer Ph.D.s annually in science and engineering. The disappointing thing is that engineering faculties have not responded with as many proposals as the science faculties. I strongly recommend that a survey study determine quickly why this is the case and how it can be remedied. There is still another approach to distributing the benefits of the ERCs to more colleges than will qualify to host them. As part of their proposals, host institutions could suggest creative ways of involving other, less for- tunate colleges-for instance, through faculty summer assignments, sab- batical leaves, student interning arrangements (both graduate and undergraduate), research subcontracts, brainstorming sessions, seminars, and review sessions. Certainly the most appropriate means to distribute benefits will depend on many things, such as area of research, geographic factors, laboratory space and equipment, and the areas of competence of faculty and students. Each situation will be different, and each will require different methods. NEW FACTORS AFFECTING ENGINEERING EDUCATION To sum up the considerations involved in the relationship between education and research, it is desirable to list those factors which are either new or have changed in importance in the last few years. 1. The breadth, depth, rate of change, sophistication, and importance of technology and engineering methods in industrial and governmental activities have created a new world for educators to deal with. To design

OCR for page 59
JERRIER A. HADDAD 135 a curriculum for today's engineering student that is as complete as that of two or three decades ago, and to keep it within four years, is difficult to the point of impossibility. 2. Engineering jobs in industry are highly diverse. The job categories of all employed engineers break down as follows (NRC, 1985b): Research Development (including design) R&D management Other management Teaching Production or inspection Other (consulting, computing, etc.) 4.7% 27.9% 8.7% 19.3% 2.1% 16.6% 20.7% Into this broad range of specialties must be factored an industrial specialty and a basic discipline. Educators cannot possibly be expert in all these activities. How to provide an education for these activities on top of an already crowded four-year program is mind boggling. Yet the increasing sophistication and importance of these activities decree that the education system somehow must accommodate them to a greater degree than before. 3. About 200 engineering colleges are predominately undergraduate institutions that produce half the B.S.s in engineering annually. These institutions lack the advantages that research institutions enjoy: world- class faculty, state-of-the-art laboratories and equipment, and the sup- porting infrastructure that these things bring. If we are to raise the quality and ability of the graduating engineer, we must focus on the graduates of these institutions as well as the research institutions. We cannot and should not aspire to make all engineering colleges into world-class research in- stitutions. However, we cannot stop short of improving the education experience for all engineering students in areas that count most. 4. Engineering in government and industry is becoming increasingly sophisticated in order to compete in an increasingly competitive world. Practitioners need to know the latest research findings, and researchers need to know the obstacles to engineering progress. Industrial concerns, except for the very large and affluent, cannot possibly do research in all the technical areas of importance to them. Consequently, as a nation we should use our government funds and education system to ensure that we appropriately cover the areas of research that are important to our success in the global marketplace. ---I 5. In the effort to hold the undergraduate engineering program to a nominal four years, courses in practical skills have had to all but disappear. The increase of engineering research on our campuses should involve as much student interning as possible so as to expose as many students as

OCR for page 59
136 RELATIONSHIP BE - EEN ENGINEERING EDUCATION ED RESEARCH possible to the real world of laboratory work. Such laboratory exposure is considerably better than the standard laboratory course. 6. A serious and perennial problem for faculty is to keep abreast of progress in engineering around the world. This makes faculty contact with practitioners essential. Any program that brings serious practitioners to the campus for technical dialogue is invaluable to the education process. Since an ERC cannot be expected to stimulate this type of interaction outside its technical area, special efforts should be made to introduce the industry people to other elements of the engineering college. Taken together, these points say that the ERCs represent an idea whose time has come. The ERCs are the first really creative response to a number of interrelated problems. We should labor hard to make them work. REFERENCES National Research Council (NRC). 1985a. Engineering education and practice in the United States: Foundations of our techno-economic future. Report of the Committee on the Education and Utilization of the Engineer. Washington, D.C.: National Academy Press. National Research Council (NRC). 1985b. Engineering employment characteristics. Report of the Panel on Engineering Employment Characteristics, Committee on the Education and Utilization of the Engineer. Washington, D.C.: National Academy Press.