National Academies Press: OpenBook

Directions in Engineering Research: An Assessment of Opportunities and Needs (1987)

Chapter: 5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview

« Previous: 4. Energy, Mineral, and Environmental Systems Research in the United States: An Overview
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 182
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 183
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 184
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 185
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 186
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 187
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 188
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 189
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 190
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 191
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 192
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 193
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 194
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 195
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 196
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 197
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 198
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 199
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 200
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 201
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 202
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 203
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 204
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 205
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 206
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 207
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 208
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 209
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 210
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 211
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 212
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 213
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 214
Suggested Citation:"5. Information, Communication, Computation, and Control Systems Research in the United States: An Overview." National Research Council. 1987. Directions in Engineering Research: An Assessment of Opportunities and Needs. Washington, DC: The National Academies Press. doi: 10.17226/1035.
×
Page 215

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Information, Communication, Computation, and Control Systems Research in the United States: An Overview Executive Summary In today's world, information is the key to a successful technol- ogy-based society. The speed and effectiveness of information communication, processing, and use are critical to the success or failure—of our national economic enterprise. Entire industries, such as banking, insurance, and law, have become dependent on computer data bases. Airline ticketing offices and the stock ex- changes can handle an enormous volume of transactions with the use of modern information processing technology. Information, communications, computation, and control (iC3) technologies are also crucial elements of our national defense, both in its manage- ment and its weapons systems. Because it is so pervasive, this area of engineering can be considered basic to all others as the world moves into the information age. That pervasiveness means that iC3 technologies are now the focus of intense competition among nations for technological lead- ership and the domination of the enormous commercial markets. Indeed, no battlefield in the struggle for international competitive superiority in technology is as strategic as this one. It is an area in which the United States has led the world, but in which we are being severely challenged by foreign competition. Our eco- nomic survival as the leading technological nation depends on our performance in key areas of ~C3 technologies. 182

IC3 SYSTEMS RESEARCH 183 At the center of these information-based technologies is the concept of an "information system. An information system con- sists of an input, at which information is gathered; a processing section, at which computation takes place; and communication of the information from input to processor and from processor to a point of control, an output, or both. Such systems are ba- sic to all intelligent processes. The engineering research problems they represent have grown enormously as our ability to transmit and process greater quantities of information per unit time has increased, making possible orders-of-magnitude greater complex- ity and productivity in every area of information processing and control. Information systems consist of hardware (i.e., sensors, actu- ators, operating devices, and other subsystems in their physical embodiments) and software (i.e., the set of instructions that gov- ern system or subsystem operation). System design requires that the functioning and interactions of both the hardware and the software be understood, and that these elements be efficiently in- tegrated and implemented. In general, the trend has been for information processing hardware to decrease in size, weight, and cost while becoming faster and more powerful. At the same time, software has tended to become more complex. Although there has been a rapid increase in the complexity and performance of hard- ware over roughly the past two decades, our ability to design and produce software has improved less rapidly. Greater productivity in software development needs to be emphasized in our national engineering research priorities, along with strong support for con- tinued increases in hardware performance. RECOMMENDATIONS 1. The speed of electronic devices has increased, whereas size, power dissipation, and cost have decreased dramatically over the past two decades. We recommend that research into materials, processing, circuits, and interconnection technologies for lC3 de- vices and components be given high priority so that the nation can maintain and strengthen its competence and leadership in these areas.

184 DIRECTIONS IN ENGINEERING RESEARCH 2. The complexity of computer-based systems has increased because of advances in technology and escalating end-user re- quirements. Increasingly sophisticated architectures and operat- ing systems are necessary to make such computer systems operate effectively. We recommend that increasing the power of complex systems and the productivity of software development for them be set as national engineering priorities. This will require substantial progress in architectures (especially distributed and parallel pro- cessing), in the integration of software, in data base management, and in large-scale communication networks. 3. The versatility and end-user friendliness of lC3 systems can be greatly improved. We recommend increased attention to both input devices (e.g., sensors, text and speech recognition transduc- ers) and output devices (e.g., graphics displays, speech synthesiz- ers, robotics manipulators) so that the true potential of advances in IC3 can be realized by humans as well as by technological sys- tems. 4. Government, industry, and universities should address the inadequacies of present engineering research facilities and equip- ment in universities, so that future practitioners and faculty in ~C3 technologies can be educated to fulfill a critical national need. 5. Universities should evaluate their organizational structures and their reward systems so that more cross-disciplinary work can flourish in both engineering research and teaching. Introduction No battlefield in the struggle for international competitive su- periority is as strategic as that of information, communications, computation, and control (DECO. These disciplines are at the root of our national strength and health in most of the key components of industry and national defense. Furthermore, they are areas in which U.S. leadership is seriously challenged by international competition, mainly from the Far East. We must exert our best efforts to maintain high-technology leadership where we have it, and to regain whatever leadership we have lost. This is a matter of national urgency; the outcome of the competition is by no means decided. If the United States is to remain economically, militarily,

IC3 SYSTEMS RESEARCH 185 and politically strong, information has to be more quickly com- municated to the point of decision and more quickly assirn~lated. To accomplish that, the country must place a high priority on key parts of these areas of engineering. Each of the disciplines of ~C3 contains both a hardware tech- nology component and a systems and software component. We must have vital engineering research programs in each of these areas. It is a fallacy to believe that we can emphasize either sector and leave the other to the competition. Software and hardware most strongly interact precisely at the leading edge of technol- ogy development. That is where advances in one sector can be capitalized on most rapidly by the other in the creation of new capabilities and new systems. The United States has been a world leader in integrated cir- cuits and other aspects of hardware technology, as well as in soft- ware systems design. It is of critical importance not only that we continue to lead in these two areas, but also that we more effectively integrate progress here. The panel notes that academia, which is funded for research primarily by government agencies in these areas, is not always able to find equipment and funds for some types of research that are important to industry. For example, bipolar integrated circuits are the foundation of most high-speed, large-scale computers; however, metal oxide semiconductor-type integrated circuits predominate in university research and teach- ing. Matching that portion of university research that is directed toward future national needs with a more clearly defined percep- tion of what those needs are likely to be is a continuing challenge. Finally, advances in ~C3 technology are crucial to several other important areas of engineering. Manufacturing now depends on robotics and automation, which in turn require a host of advances in effective and reliable computation, communication, and control. Engineering design of a vast range of key products and systems- such as computer chips, automobiles, buildings, and industrial processing depends critically on ~C3. Therefore, it is essential that we consider the measures that can be taken to strengthen engineering research in these areas. Given the importance and impact of these technologies on defense, in industry and commerce, and in the lives of individual citizens, coupled with the strong research base that their advance- ment requires, it is not surprising that government support and policies play an important role in determining the pace and vigor

186 DIRECTIONS IN ENGINEERING RESEARCH of research in these fields. Thus, in addition to identifying key research needs, the pane} also examines federal policy toward re- search in ~C3 technologies as well asthose policies that affect the conduct of engineering research in general. In a concluding section, special topics relating to the overall health (both present and future) of research in these areas are addressed. Particular attention is paid to the working environment for university faculty in lC3 and to the adequacy, both quantitative and qualitative, of graduating engineers who form the talent pool for research. Research Needs The Most Import ant Areas of ~formation, Communications, Computation, and Control Systems Research In view of the critical importance of research advances in this field to our national competitiveness and security, this section of the report is the primary focus of the panel's work. The pane! considered more than 70 distinct research topics in the broad categories of research encompassed by its scope of concern. In the course of its evaluation, the pane! took into account the opinions expressed by engineering deans and faculty, selected researchers, and various professional engineering organizations in response to a survey conducted! by the Engineering Research Board (see the Appendix). The results of this assessment are discussed in the following pages. The discussion is organized In terms of (1) the hardware and (2) the software and systems research needs associated with ~C3 technologies. It is important to bear in mind the point made in the introduction: that hardware and software/algorithms in ~C3 are two indispensable sides of the same coin and must be supported equally. Across the spectrum of engineering research needs identified here, there is an urgency with respect to national interests that must not be ignored. A number of the identifiecl research topics reflect the panel's perception that the use of information processing systems is often Innited by the input of real-worId (i.e., analog) information into

IC3 SYSTEMS RESEARCH 187 the systems, and by the output of information in a form that is usable either for control or for human interpretation. Thus, input and output devices ant] their associated software and algorithms are a bottleneck in the full application of systems whose processing capabilities have, at present, exceeded their general utility. HARDWARE ELEMENTS IC3 systems are constructed from hardware elements. Rapid improvements in the function, density, and performance of these hardware elements have been the driving force behind the revo- Jution in electronics. Continued progress requires ever-increasing sophistication in our ability to control the materials and processes required for component fabrication. This is extraordinarily fertile ground for the expansion of engineering research to solve problems relating to international technological competitiveness. The issue of how to achieve vastly improved manufacturing techniques for the production of lC3 hardware, especially com- puter and communications devices, is very important. A detailed discussion of IC3-related manufacturing research needs is outside the scope of this report. These issues are addressed in general in the report of the Pane! on Manufacturing Systems Research of the Engineering Research Board. Some of the issues specific to the fabrication of computer devices are addressed in the report of the board's Panel on Materials Systems Research. (Both reports are in this volume.) COMPUTER DEVICES Several categories of devices underlie progress in computation; these are integrated circuits, interconnection structures (so-called ~packaging"), and magnetic and optical storage. In addition, the need for improved methodologies for testing these devices is be- coming very pressing. Continued progress in theoretical and ex- perimental research in each of these areas is vital to the future health of the U.S. computer industry. (See also the discussion of "Semiconducting and Magnetic Materials in the report of the Panel on Materials Systems Research.)

188 DIRECTIONS IN ENGINEERING RESEARCH Integrated Circuits Progress in integrated circuits for computers takes place in two principal directions: logic circuits and memory circuits. In the case of memory, the principal thrust (as is well known) is increasing density, as measured by the number of bits stored per chip. In the case of logic chips, there are again two principal thrusts: high performance, as measured by circuit switching speed, and high density, as measured by total number of logic circuits per chip. Very large-scale integration (VSLI) has been applied to mem- ory and to logic circuitry. However, none of the very large-scale, general-purpose computers, and few engineering-scientific super- computers, rely on metal oxide semiconductor VESI for their crit- ical circuitry. Instead, the hundreds of thousands of logic circuits in these economically and scientifically important computers are almost exclusively bipolar silicon logic circuits. These circuits are much faster than microprocessors, although they are of substan- tially lower circuit density, and are usually the best means for pro- ducing the desired high performance. The continued improvement and enhancement of bipolar silicon logic is of great significance to the future well-being of U.S. computer technology. In the minds of some, its significance wit' be as "Teat as the continued improvement in VI S] memory and logic density, size, and performance. In both VLSI and bipolar technologies there is substantial engineering re- search to be done; however, the pane! notes the remarkably small amount of attention given in universities to the technology that Is fundamental to the present and future success of large-scare, general-purpose and scientific computing, namely, silicon bipolar devices, processes, and chips. International competition in this area is very keen, with great strides being made in Japan. Nigh-Density Structures and Fabrication Two advances in semiconductor devices are critical to con- tinued performance improvements of integrated circuit-based sys- tems: (1J advances in submicrometer device structures and fabrica- tion methods and (2) advances in three-dimensional (~-D) devices and CiTCUits. Both of these research efforts seek to continue the growth in complexity at the integrated-circuit chip level, while im- proving speed and function. Both tasks pose substantial challenges for industrial and university laboratories alike. The fabrication of submicrometer structures requires special process equipment

IC3 SYSTEMS RESEARCH 189 whose registration and resolution are capable of dimensions less than the wavelength of visible light. Similarly, successful 3-D structure processing* requires sophisticated equipment capable of depositing defect-free materials. Packaging and Interconnection Technology Because large-scale, general-purpose and scientific computers require so many high-performance, relatively low-density chips, the "package" that interconnects their hundreds of signal and power connections is of crucial importance to achieving desired system performance. This desired performance involves fast signal propagation, rapid heat removal, and highly reliable mechanical and electrical interconnection with the next level of the packaging hierarchy. Improvements in signal delay, power dissipation, and parasitic coupling all require more sophisticated technology than is now avait- able. Although these have not traditionally been v dewed as techni- cally interesting problems, their solution is essential to achieving the highest possible system performance. Ceramic and polymer engineering, thin-fiIm metallurgy, the mechanical properties of composite structures, and optical interconnection are all relevant disciplines here, and all are areas in which university research can and should be vigorous. Novel approaches to package design are of potentially high leverage, and much more innovation is needed. Magnetic and Optical Storage With regard to information storage media and devices, there are a number of promising approaches that can increase storage density by several orders of magnitude. For magnetic storage, both vertical recording and signal encoding techniques offer significant improvements over the current state of the art. Vertical recording offers a means for packing magnetic transitions more densely; signal encoding techniques offer a way of using those transitions to store information more efficiently. Run-length coding is one form of signal encoding in use today that achieves about two to three times the storage capacity of unencoded data. Yet this gain is small compared to what is theoretically achievable. *3-D devices have circuit elements stacked vertically to conserve space on a chip.

190 DIRECTIONS IN ENGINEERING RESEARCH Optical storage technology has developed to the point that it could potentially supplement magnetic storage for auxiliary memory. However, erasing an] retrieval of store] information is not possible for some optical storage devices, and for others it is far slower and more costly today than for magnetic storage systems. Research into optical storage may be able to eliminate these and other drawbacks of the technology so as to create storage svatems superior to magnetic storage. As in many engineering research fields, the application of computer-aided design (CAD) technology is crucial in improving the productivity and quality of all forms of computer logic and memory circuit design. Engineering research on new design tools ant} methodologies has played and will continue to play a vital role in this aspect of the technology. . ~ , Hardware and Subsystem Testing CAD is also of great potential value for testing devices and subsystem assemblies. The chips and packages of present and future computer systems are so complex that their testing and simulation now consumes a large amount of time and adds sum sta~tially to their cost. Even the methodology for economical testing of large-scale logic chips and systems is poorly understood and needs further work. Testing should be able to be accomplished on three levels: (1) the design phase (i.e., is it a good design?), (2) the chip-manufacturing phase (i.e., does each chip work?), and (3) the in-operation phase (i.e., self-testing by the chip to ensure that it is functioning correctly). New types of testing methods, such as "contactless" Abeam and laser-assisted testing, are needed. These methods will require considerable research. COMMUNICATIONS The devices and components that support communication also offer substantial challenges to engineering research. Device re- search for optical communication is a very important area, one in which the Japanese are very strong and in which there is enormous commercial potential. Devices that permit the switching of light signals from one transmission channel to another are emerging as an important area of research; this switching technology cur- rently requires more attention than does the technology of optical

IC3 SYSTEMS RESEARCH 191 transmission, which is being well researched. Research on devices that efficiently transform electrical signals into optical signals, and vice versa (i.e., lasers and detectors), is of great technological importance in communications. Particularly promising materials include both the ITI-V and IT-V! compound semiconductors. Im- portant and novel structures include superIattices and quantum wells, in which a strong interaction between optical and electrical signals can be obtained. In order to build the all-optical communications systems of the future, a number of devices familiar in radio will need to be in- vented in the optical domain. Such optoelectronic devices include amplifiers, filters, isolators, multiplexers, and switches. These op- tical "plumbing" building blocks will enable the implementation of multichannel communication systems similar to those found in radio transmission, except that the capacities of the optical sys- tems will be enormously greater than their radio counterparts. After these optical devices have evolved, still another generation of development and research waif be required to integrate them onto rn~crochips in the manner in which electronic circuitry Is now produced. SENSORS FOR CONTROL OF SYSTEMS A great deal of progress has been made in general computa- tion; however, not enough progress has been made in getting infor- mation to the computer. The hardware elements that gather data are sensors. These input devices translate temperature, move- ment, thickness, flow, and many other physical parameters into electrical signals that may be used in communication, computa- tion, and control applications for example, in automation of the total manufacturing process. There is a great need for improved sensitivity, linearity, resolution, wavelength response, and degree of miniaturization and integration of sensors, as well as for sensors for parameters not now amenable to sensing (e.g., gas phase chem- ical composition or the fidelity of a manufactured part to desired specifications). We note, moreover, that real-t~me control is de- pendent on the ability to rapidly and accurately sense process and control variables, so that in certain applications it is the sensor and/or actuator technology that is limiting further progress. The development of sensors is important for improving control of large systems in which the placement, location, and number of sensors

192 DIRECTIONS IN ENGINEERING RESEARCH is a critical matter. Thus, research on sensors and actuators is greatly needed. SYSTEM ARCHITECTURE, ALGORITHMS, AND SOFTWARE The foregoing discussion dealt with hardware-oriented re- search needs in IC3. To make those devices and components functional that is, to give functionality to the systems in which they are embedded requires corresponding advances in systems architecture, algorithms, and software. Indeed, many of the most difficult problems today in the IC3 field are in designing special computer architectures, in producing enormously large and com- plex software programs, and in evolving efficient computational algorithms for problems of overwhelming size. COMMUNICATIONS Many U.S. researchers are now working on problems in the communication field. However, certain problems still require high- priority attention. For example, one of the largest software efforts in coming decades will be the development of a system underlying the national communication network. This network will eventually overlay the current telephone network with a fiber-optics-based network combining voice, data, and video transmission. The ma- jority of transmissions will be in a digital format. Switching will be accomplished by the time-interchange of bits within transmitted data streams, and by the routing and storage of packets of infor- mation containing address headers. The need for traffic and flow analysis of these streams of packets, and for efficient and reliable protocols to manage the interchange of data, has grown in recent years. As long-haul communications have become more and more efficient, much of the research interest has turned toward the bot- tIenecks in local distribution and collection. Networks of users (generally computer terminals) within the area of a building, a campus, or an industrial complex require methodologies for shar- ing a broadband medium such as a coaxial cable or fiber. Such a network is known as a local area network. On the next higher level, the breakup of the Bell System and the evolution of a range of choices for communications access have led to increased interest

IC3 SYSTEMS RESEARCH 193 in metropolitan area networks to link large buildings and multi- ple users in an efficient manner to the long-distance network. In all of these problems, computer software and algorithms play a key role. Some of the largest computer programs, requiring hun- dreds of man-years of development effort, are required to direct the switching activity of a modern telecommunications node. COMPUTER SOFTWARE Large software systems (for example, the operating system of a large multiprocessor) are some of the most complex creations ever devised. Our ability to design, code, test, and modify large software systems has improved somewhat in the past decade. Yet the need for such systems is outstripping our current capabilities. A sustained basic research program on methodologies for the effl- cient development of large software systems is sorely needed. One thing that would enhance software productivity would be the abil- ity to integrate software subsystems. Methodologies are needed by which software can be made readily integrable with other software. Also important is research on the compatibility, reuse, and stan- dardization of key software modules (e.g., floating point modules, encryption algorithms, and communication protocols). An issue that becomes increasingly important with the growth in the size of software systems is the general issue of reliability, testing, and verification. Continuing attention needs to be paid to this issue in order to combat the combinatorial explosion of testing sequences that occurs as systems get larger and more complex. There has been much discussion about distributed computer systems in recent years. Yet many of the basic issues have not been properly addressed. These issues would benefit from increased attention by academic researchers; this would complement the intense interest in distributed processing in industry. Communi- cations networks currently permit computers in different places of the same or different architectures and software environments to "talks to each other at a relatively low level. What is needed is a way of linking such computers and their associated data bases with each other and with large, central data bases so that a wide array of services is performed in a manner that is both efficient and transparent to the user. Some types of sensors (e.g., those mounted on satellites) gen- erate enormous amounts of data for input to processing systems.

194 DIRECTIONS IN ENGINEERING RESEARCH Additional work is needed on the development of techniques, such as data bases and signal processing algorithms, that will permit reai-time processing of data generated in such large volumes. PARALLEL COMPUTATION Within this century, improvements in the technology underly- ing computing can be expected to yield increases in speed of one or two orders of magnitude. In certain applications, the demands for computing power are far greater than are likely to be met by speed improvements in components and by evolutionary improve- ments of the basic (von Neumann) computer architecture. Certain approaches that have the potential for at least a thousand-fold im- provement in speed should be pursued. The most promising areas for research are in architectures for parallel computation. Special Purpose Parallel Architectures This area of real-time processing often lends itself to very specialized architectures (e.g., systolic arrays) that promise great speed at relatively low cost. Similarly, there are great advantages to specialized parallel architectures for processing visual inputs. The complex memory structures used in artificial intelligence lend themselves to computer organizations containing a million or more nodes that interact in parallel. Parallel Architectures for Numerical Computation Current supercomputers are usually based on machines that can manipulate vectors in parallel. Such designs can yield one or two orders of magnitude of additional speed, especially by increas- ing the number of vectors that can be manipulated simultaneously. The NSF's program on supercomputing will give the academic community access to current and next-generation machines. This opportunity should Reprove our ability to design algorithms that effectively use the available parallelism. In particular, increased attention needs to be paid to redesigning the numerical algorithms for parallel architectures. Attention is beginning to be paid to non-vector-oriented par- alle! architectures for numerical computation. Such architectures appear to be needled for solving physical problems (e.g., weather

IC3 SYSTEMS RESEARCH 195 prediction, fusion machine analysis, and viscous fluid flows) in their full 3-D form and within a reasonable length of time. General Purpose Parallel Architectures The greatest risk, and possibly the biggest payoff, involves parallel architectures that are intended for relatively general ap- plication (e.g., data flow computers). It is not clear whether there are many problems that lend themselves to such massive paral- lelism. It is also not clear whether the architectural designs can deliver the promised speeds. One of the major applications for these computers is rule- based expert systems (see the following section). The development of such a system is one of the goals of the Japanese Fifth Gen- eration Project. Such architectures are also being investigated in Europe and the United States (viz., the Strategic Computing ef- fort of the Defense Advanced Research Project Agency DARPA). Clearly, much could be learned from the continued analysis and experimental contraction of such machines. MAN-MACHINE INTERACTIONS AND ARTIFICIAL INTELLIGENCE The interface between the analog and digital worlds is seen most distinctly in man-machine interactions. This interface is often not very "friendly," and it is here that some of the most notable bottlenecks exist in our attempt to bring computing power into our everyday lives for practical uses. We discuss three areas of needed research next. Rule-Based Expert Systems Recently, there has been much interest in applying the tech- nology of rule-based expert systems to the solution of a wide range of problems. We are somewhat concerned that the approach has been oversold. Nevertheless, this concept has broadened, in a very real way, the general idea of what computers can do. Namely, there is a growing realization is being widely realized that com- puters, like people, can be quite useful in areas in which a complete algorithm does not exist. Increasing the number of rules often degrades the speed of rule-based systems. This problem is being addressed by designing

196 DIRECTIONS IN ENGINEERING RESEARCH massively parallel computers, as noted previously. Increasing the number of rules can also degrade the effectiveness of rule-based systems because of the complex interactions of new rules with existing ones. This problem is being addressed by restructuring rules in a variety of ways so that the complexity of interactions is reduced. Knowledge-Based Systems A rule-based expert system is a specific type of "knowledge- based~ system. The first knowledge-based systems were large systems such as MACSYMA, for mathematical formula manipu- lations, and Dendral, for analyzing molecular spectra. Such large knowledge-based systems can be characterized as requiring long- term basic research into a particular problem area (e.g., medical diagnosis) in which a key difficulty is the representation of the large amount of knowledge that humans have about the given area. Additional basic research on the particular input problem of representation is clearly needed, as are further examples of large, knowledge-based systems. Natural Language Understanding If computers had the capability to understand typed queries in natural language, it would greatly facilitate and improve the interaction between man and machine computers would be far more "user friendly." Natural language understanding presents very challenging long-term research problems. Much progress has already been made on the syntax of natural language. Most of the challenge now lies in the area of semantics and discourse. ROBOTICS AND AUTOMATION Robotics and automation present exciting opportunities and major challenges for lC3 systems research. This is a prime exam- ple of the need to improve both the input and output of data. Coordination of multiple flexible manipulators, coordinated con- tro! of multiple robots, and automation in manufacturing require complex decision making and real-time control involving enormous communications and computations. Integrated research in com- puter vision, artificial intelligence, decision and control theory,

IC3 SYSTEMS RESEARCH 197 large-scale system theory, algorithms for decentralized computa- tion, and data communication should lead to major advances in robotics and automation. The emphasis here is on the integration and merger of the separate research areas into a unified approach to problems in robotics and automation. For further discussion relating to the applicability of robotics and automation to the manufacturing process, see the report of the Pane! on Manufacturing Systems Research. CONTROL SYSTEMS The development, planning, design, and operation of systems for communication, computation, defense, transportation, etc., necessarily involve the issue of real-time control. Control is con- cerned with the determination of decision variables on the basis of feedback signals from sensors and other data,gathering instru- ments in such a way as to meet system objectives. Real-time computation and data processing are necessary to produce de- cision variables from sensor data. Complex, large-scale systems involve a large number of decision variables and a great many sys- tem variables whose status is monitored by sensors at a high data rate. It is not always possible or desirable to centralize information gathering and the real-time computation of decision variables. One of the most critical needs in control systems research is the investigation of decentralized control strategies in which local deci- sion variables are determined on the basis of local sensor signal. As mentioned earlier, with large-scale systems the determination of what system variables should be measured and the proper number and placement of sensors to achieve optimum overall effectiveness are both extremely important. Because local control based on lo- cal sensor data ignores data gathered elsewhere, local perceptions of system behavior may differ. Furthermore, local decision vari- ables generally affect system variables elsewhere. Thus, each set of local inputs and outputs leads to a different mode! of the same system. There is a need for a theory underlying the behavior of large- scate systems. Research is needed to determine what computa- tional aigorithrns are necessary for decentralized control of large- scale systems. Multimodeling, multicriteria optirn~zation, stabi- lization, and coordination are some fundamental issues in decen- tralized control.

198 DIRECTIONS IN ENGINEERING RESEARCH Policy Issues on Federal Support of Research SCOPE This section of the report focuses on the federal government's role in, and influence on, fundamental engineering research in {C3. Such research is performed in universities, national and federal lab- oratories, industry, and cooperative or collaborative institutions. Federal policies have a great impact on the vitality, effectiveness, and potential usefulness of such research. The federal government's role includes . Direct support by the NSF, whose unique and specific re- sponsibility is support of fundamental research in universities and related research centers, as well as support of graduate education. . Direct support by mission agencies such as the Department of Defense (DOD), NASA, etc. . Encouragement, through tax incentives or other means, of research performed by industry (both by individual companies and consortia). ~ Encouragement, via tax incentives and other means, for industry-university cooperative research and education. ISSUES THE NSF'S ROLE IN SUPPORT OF ACADEMIC RESEARCH IN IC3 SYSTEMS ENGINEERING IC3 system technologies have made phenomenal progress over the last two decades. Much of this advance has been based on outstanding research performed in industrial laboratories, where the necessary state-of-the-art experimental facilities and computa- tional capabilities are available. Nevertheless, academic engineer- ing research is absolutely essential for the future health of these cutting-edge technologies. Electronic and information engineering research is closely tied to advances in the physical, chemical, and mathematical sciences. One need only review the remarkable progress made in advanc- ing these technologies in organizations such as the Bell Telephone

IC3 SYSTEMS RESEARCH 199 Laboratories, the IBM Research Laboratories, and several out- sta~ding university research centers to appreciate the value of coupling fundamental science and engineering research. Syner- gism between science and technology benefits both, and is one of the great strengths of the American science and engineering enterprise. Although industry performs about 75 percent of the R&D in the United States, it performs a much smaller fraction (15 to 20 percent) of the basic research (National Science Foundation, 1983~. The structure of American universities permits the interaction of electronic and information engineering research with its scientific and mathematical base. Therefore, it is important to continue to enhance academic research in these forefront engineering fields and to encourage interdisciplinary research programs that combine the appropriate science and engineering disciplines. It must also be recognized that increasing competitive pressures in communica- tions and computation tend to force industry into a search for near-term results; we must Took to the universities for much of the research that will result in the long-term, unforeseen advances in these fields. The NSF, which is responsible for supporting much of the science and engineering carried out in universities, is therefore a key agency for maintaining and strengthening ~C3 engineering research. (See, for example, Farber, 1985.) The success of the engineering research enterprise is also highly dependent on the availability of a pool of doctoral-level profes- sionals. Responsibility for graduate education rests with those universities that have strong academic research programs. The interaction of research faculty and graduate students is essential to the research enterprise, and provides the best environment for training graduate students for careers in research. The NSF and the federal mission agencies are major supporters of both academic research and graduate training through fellowships ant] research assistantships and through grants and contracts. They must con- tinue to fill this important role. The new NSF initiative to establish multidisciplinary Engi- neering Research Centers (ERCs) on university campuses can be- come a very effective supplement to its basic individual-investigator grant system. In the first year of the program, NSF established six ERCs involving eight universities. These six facilities are to receive $94.5 million over the next 5 years and are likely to receive significant industrial support as well.

200 DIRECTIONS IN ENGINEERING RESEARCH Three of the six initial centers are in areas relevant to ~C3 SyS- tems research. The Center for Robotic Systems in Microelectronics at the University of California at Santa Barbara will investigate new automation technologies for the fabrication of sem~conduc- tor devices. The Engineering Research Center for Telecommu- nications at Columbia University will focus on communications networks to integrate various communications transmissions. The Center on Systems Research at the University of Maryland, in collaboration with Harvard, will conduct research on the applica- tion of artificial intelligence and VESI to automatic control and communications systems. With the addition of new ERCs, the tote] yearly budget of the Engineering Research Center Program is planned to reach $100 million in federal funds and is expected to attract substantial addi- tional sums from private industry. The program will undoubtedly have a significant impact on the health and vitality of engineering research in the university environment. THE DOD'sRoLE Since World War IT, the DOD has been a dominant source of support for basic and applied engineering research. Today more than ever, IC3 system provide the innovative and advanced technologies that are at the heart of our modern military sys- tems. Command and control of our military forces, revolutionary advances in weapons system, worldwide secure communication systems, and surveillance systems are all critically dependent on modern electronics, computers, sensors, and control systems. It is therefore not surprising that the DOD has been the mission agency with the greatest need for (and with the largest support for) basic and applied research in the technical areas of interest to this panel. The DOD's budget for R&D Is divided into five accounts, two of which support fundamental engineering research. Basic and applied research, the 6.1 account in DOD parlance, sustains the science and technology base, whereas 6.2, the exploratory develop- ment account, also contains some elements of fairly fundamental research such as the architecture of a new generation of super- speed computers. In FY84, universities received 50 percent of the 6.1 basic re- search funding. Federal Contract Research Centers received 20

IC3 SYSTEMS RESEARCH 201 percent, and government laboratories received 30 percent. DOD basic research funds are managed by a number of organizations, including the Office of Naval Research, the Army Research Of- fice, the Air Force Office of Scientific Research, DARPA, and the Office of the Deputy Undersecretary of Defense for Research and Advanced Technology, as well as the various DOD laboratories. About one-fourth of the 1984 6.1 funding for electronics research was managed by DARPA, along with some 45 percent of the fund- ing for mathematics and computer research. DARPA's program in support of computer research, computer communications, and artificial intelligence has been critically important for progress in these fields. Similarly, the Joint Services Electronics Program has been a major factor in the establishment of outstanding research centers at many major universities. Although this program has expanded somewhat, over the past decade, it has become propor- tionally less significant in the overall engineering research scene. However, the FY86 DOD budget indicates a renewed realiza- tion of the need to enhance engineering research as an essential el- ement in our national security. The FY86 defense budget contains a $25 million program to support high-risk basic research, multi- disciplinary centers, research equipment, and research fellowships. There are other initiatives as well. An example is the Software En- gineering Institute, established by the DOD at Carnegie-Mellon University for research on the production of software for defense applications. The DOD plans to provide $103 million for this institute over the next 5 years. The research performed by the institute is expected to be unclassified and in the public domain. This recognition by the DOD that the vitality and magnitude of fundamental engineering research and education are essential to the national welfare and defense is welcome. The DOD initiatives will sustain the needed growth of university research; they are indicative of a renewal of emphasis, on the part of DOD, on the long-term health of university research rather than on short-term defense requirements. The pane! hopes to see this enlightened outlook and support continue. There is some concern that continuing attempts by certain re- sponsible and dedicated elements within DOD to restrict the dis- semination of unclassified research at open meetings or in journal publications ~ the United States might impede the essential flow of internal communications, thus hampering the overall health of en- gineering research as well as DOD's harmonious relationship with

202 DIRECTIONS IN ENGINEERING RESEARCH the universities. Clearly understood and implementable guidelines must be established and adhered to by all concerned if we are to ensure that research results benefit the technical community while satisfying the legitimate needs of national security. Recommendation We recommend continuing the current system of support of fundamental engineering research by the various organizational elements of the DOD. We welcome DOD's new initiatives that en- hance the vitality and funding of fundamental engineering research at the universities. We alto want to emphasize that the effective- ness of university research lies in the investigation of longer term fundamental issues rather than in meeting the shorter term defense requirements that are the proper province of private industry. We also believe that any restriction on the dissemination of unciassi- fied university research results may be detrimental to the research enterprise, and thus to the longer term advancement of defense programs. THE FEDERAL ROLE IN ENCOURAGING RESEARCH IN INDUSTRY Although universities have the intellectual resources, the en- vironment, and the incentives to lead in fundamental research, a great many of the engineering research breakthroughs in lC3 have actually occurred in industrial laboratories. It is essential for the health of innovation in electronics and computers that government policies encourage research in the industrial sector. A number of very significant industrial research activities can be attributed to favorable government policies. For example, most of the major DOD contractors maintain research laboratories that conduct competent research of great value to DOD agencies. A very significant and perhaps dominant portion of this research is initiated by the industrial research team and is partially paid for as an allowable overhead cost. It is desig- nated Independent Research and Development (IRAD). Because TRAD efforts are funded on a cost-sharing basis, high-risk, high- payoff areas can be chosen for investigation. Furthermore, IRAD may be essential for maintaining the continuity of R&D funding through the vagaries of DOD budgetary fluctuations.

IC3 SYSTEMS RESEARCH 203 Congressional tax acts have provided a number of tax incen- tives to encourage industry to conduct R&D. A substantial portion (perhaps 10 percent) of the increased industrial R&D funds are used to conduct fundamental engineering research. Recognizing the effectiveness of industrial consortia developed in foreign countries, the federal government has permitted indus- try, within the constraints of the antitrust laws, to form research organizations supported by companies in the same industry. The Microelectronics and Computer Technology Corporation (MCC), established in 1983 in Austin, Texas, is a prime example of such a consortium. As development intensifies in the lC3 field, a vast array of new products and services will continue to emerge. Innovations in areas such as software are already giving rise to major alterations in the concept of intellectual property. The extension of existing laws to cover new types of property, and questions about the enforcement of those laws in an environment in which there are often few obstacles to misappropriation, are drawing serious attention in Congress and elsewhere. The lack of an adequate policy in these areas will retard industrial research in certain areas of lC3. Recommendation Fundamental research performed by industry is an essential ingredient in the progress of electronics and computer ROD. The federal government should continue to encourage industrial research through direct contractual support, through DOD IRA D programs, by means of tax incentives, and by encouraging research consortia to be organized where appropriate. INDUSTRY- UNIVERSITY COOPERATION Industry and universities have had a long and fruitful re- lationship. Academia has helped to maintain the scientific and engineering momentum that is the basis of our high-technology industry, both by providing trained manpower and by conduct- ing the research that is part of the training process. Yet, today, industrial research actually leads the universities in some areas because the facilities and staff required are beyond the ability of most universities to afford, even with federal support. It has thus become necessary to develop, with the help of government, new approaches to industry-university research collaboration.

204 DIRECTIONS IN ENGINEERING RESEARCH The NSF has established 20 industry-university cooperative research centers. The center at the University of Arizona, for ex- ample, was established to study the electro-optical properties of materials and structures for high-speed computers and communi- cations. During its first year, it received joint funding from the NSF and a number of private companies. Currently, the NSF is establishing the Engineering Research Center Program, described earlier, for which industry funding is anticipated. The Semiconductor Research Corporation (SRC), founded in 1982, has already organized centers of excellence, with major, long-term thrusts at three universities. The SRC has contracts at 32 universities and supports 300 students and faculty members. Its member companies include AT&T, Burroughs, CDC, DEC, DuPont, GE, HP, Honeywell, IBM, Intel, Motorola, RCA, TI, Union Carbide, Westinghouse, and Xerox. In addition, a number of states have initiated successful pro- grams involving joint state, university, and industrial partici- pation. The California MICRO Program to support industry- university collaboration in microelectronics research is one ex- ample. The Microelectronics Center of North Carolina, MCC in Texas, and the Center for Industrial Innovation at Rensselaer Polytechnic Institute in New York are further examples of state, university, and industry cooperation to foster the establishment of technology centers of excellence. All of these cooperative initiatives have the same underlying objectives (Office of Technology Assessment, 1985~: ~ improved research and new knowledge- mutual benefits are derived from the sharing of knowledge, funds, equipment, personnel, and technology; Or ~ education and manpower greater research activity and better facilities result in a strengthened academic program gener- ally, as well as beneficial contacts with industry from the point of view of students and faculty; and . economic growth - improvements in the research base and the technical manpower base lead to greater vigor in a region's overall economy. Properly guided, and with sufficient sustained funding and a com- mitment to these goals, these initiatives have great potential for advancing the nation's strength in the research areas upon which they focus.

IC3 SYSTEMS RESEARCH 205 Recommendations The essential character of university research in maintaining OUT leadership in high-technology industry, coupled uoith the high cost of adequate facilities and staff, requires cooperative initiatives to Upgrade the facilities, capital equipment, and salaries of univer- sity faculty and research staff. These initiatives, if they are to be sustained;, must provide direct benefits to industrial sponsors white retaining freedom of research and of publication by the universities. We commend the federal and state initiatives for establishing research centers. These research centers should complement, and not replace, investigator-initiated research. The Health of the Field: An Assessment THE HEALTH OF THE EDUCATIONAL SYSTEM In this section we examine the health of the educational system in the iC3 field. A primary focus is on the quality of life of uni- versity faculty in this field. Other important topics discussed are the status of equipment and facilities, and the cross-disciplinary research approach. We also examine the adequacy of new talent, in terms of both quality and quantity at all levels (B.S., M.S., and Ph.D.~. In particular, we comment on the continuing need to retain foreign Ph.D.s both in industry and academia. FACULTY For a variety of reasons, the working environment for univer- sity faculty in virtually every engineering discipline has declined sharply in recent years. In those disciplines most closely associated with the lC3 field, key issues include high student-to-faculty ratios; diminished ability to attract new faculty (especially recent Ph.D.s); salary levels; adequacy of research and teaching equipment;

206 DIRECTIONS IN ENGINEERING RESEARCH limited funds for new facilities; and the difficulty of attracting and keeping research staff. Despite the enormous growth in enrollments, there has been no comparable increase in the number of the faculty members in these departments. Neither is the imbalance likely to correct itself. The growing rate at which high-technology products are propagated into our society argues for a continuing demand for electrical engineering and computer science graduates in the next decade, so that student demand for these courses of study is likely to remain high. Partly because electrical engineering and com- puter science departments have been unable to fill empty faculty slots, and partly to ensure a reasonable balance in enrollments within the engineering school, many universities have had to limit enrollments in those departments (to about one-third of total en- gineering enrollment in most cases). As a result, many qualified students wishing to major in this vital field must be turned away. This pressure of students wishing to major in computer science may be lessening, but it is offset by students in different fields who wish to acquire a deeper knowledge about computing. At the same time, the attractiveness of opportunities in indus- try has reduced the number of students seeking the doctorate, as well as the number of new Ph.D.s interested in academic positions. This reduction has made it difficult to recruit the additional faculty needed to lighten the teaching and supervision load. The Institute of Electrical and Electronic Engineers (IEEE [19854) reports that openings in this field now exceed several hundred, and many more new openings would certainly appear if the existing ones were filled. The situation has tended to be exacerbated further by the departure of mid-career faculty to industry. According to a recent Office of Technology Assessment (1985) study, the rate of outflow of faculty in computer science is twice that of any other engineer- ing field. Clearly, many are leaving to join new entrepreneurial companies bringing the latest research advances to market in the form of high-tech products. (Recent downturns in the computer industry are probably stemming the outflow, however.) The pane] believes that TV- and computer-based instructional resources, used imaginatively, can reduce some of the pressure on faculty by providing an alternative way to present undergraduate engineering students with high-quality instruction. Educational technology could be particularly applicable in beginning courses

IC3 SYSTEMS RESEARCH 207 with large enrollments. With or without the greater application of educational technology, there will still be a need for more teaching assistants to leverage the efforts of faculty in electrical engineering and computer science departments. Industry/academ~a salary differentials have been a major con- tributor to the faculty shortage problem. Recognizing that fact, state legislatures and university administrators have made a de- termined effort to increase starting salaries for new faculty in these disciplines so that they might compare more favorably with starting salaries for Ph.D.s in industry. At some institutions, this corrective measure has created a serious compression of the salary structure although this is by no means a universal problem. In general, electrical engineering/computer science faculty salaries are improving and are somet~rnes very competitive. Indeed, panel members from prestigious companies report that their organiza- tions have, on occasion, found it difficult to match university offers to their high-level research employees. In addition, consulting sum stantially augments faculty income at all levels especially in IC3. Increasingly, the main problem is not industry/academia salary differentials for Ph.D.s, but simply the unattractiveness of pursu- ing a Ph.D. given the strong appeal to B.S. graduates of jobs in these fast-moving fields in industry. Recommendations University administrators should continue granting competitive salaries to faculties in electrical engineering and computer science departments. Attention must be paid to the salary structure of mid-career and senior faculty as well as to that of junior faculty. A new program of faculty fellowships should be instituted by industry and/or the federal government. Such fellowships should not require the preparation of lengthy proposals or reports. They would give faculty members flexibility in their research that is not presently available. The Presidential Young Investigator Awards program is one example of this kind of support; more such sup- port is needed, especially for more senior faculty members. These programs would be especially useful in the case of well-established researchers who wish to shift the focus of their research, perhaps to new areas of inquiry in which results would not be immediately produced. The number of fellowships available for this purpose

208 DIRECTIONS IN ENGINEERING RESEARCH would be small (a few percent of the total); but they would offer established researchers an opportunity to "break out of the mold." With regard to educational technology, the pane] supports the recommendation of the Committee on the Education and Utilization of the Engineer (CEUE [National Research Council, 19854) namely, "These tools should be applied as rapidly and as fully as practicable in all academic programs in such a way as to enhance the quality of engineering education. Engineering schools should be encouraged to create programs for development of edu- cational technology by faculty, with shared institutional, industry, and government funding. EQu~PMENT AND FACILITIES Much attention has recently been paid by industry and govern- ment to the problem of obsolete research and teaching equipment in colleges and schools of engineering (especially in undergraduate labs). The cost to modernize this equipment has been estunated to be $1.2-$2 billion and growing (Haddad, 1983~. The most se- vere problem is in those areas in which technology is advancing the fastest that is, electrical engineering and computer science. Companies have been generously responsive to this problem. For example, in 1983 IBM and Digital Equipment Corporation to- gether donated $50 million in equipment to MIT. In the same year, Hewlett-Packard donated some $22 million to universities, mostly in the form of equipment. Apple Computer, Inc., has do- nated more than $21 million in personal computers to schools at ah levels. Other contributions have been made by Wang, IBM, NCR Corporation, and Honeywell. Despite this assistance, however, the problem remains enor- mous; and it is a moving target. Adding to the situation is the fact that gifts of equipment do not involve funding for maintenance and other operating costs which can greatly exceed a university's budget for overhead expenses of this type. Advanced electronic equipment, no matter how current and valuable, is useless if it cannot be operated and adequately maintained. The state of buildings and laboratory space so-called Bricks and mortar" is a related problem for engineering departments generally. Because the federal government essentially eliminated support for construction of facilities in the 1960s, physical plants have deteriorated alarmingly. This problem is especially acute

IC3 SYSTEMS RESEARCH 209 in lC3, because the rapid changes in the field cause facilities to become obsolete very quickly. Because the state of the art in lC3 equipment changes so fast, universities face an additional problem relating to their cost accounting practices. For determining indirect costs, they use a different depreciation schedule from the accelerated depreciation currently used by industry. Equipment is depreciated over 16 23 years and capital equipment over such long periods, universities encounter a financial problem in the renewal of advanced equip- ment for research. Recommendation A tong-term program for support of both equipment and facil- ities is urgently needed). The pane! strongly supports the recom- mendation of the Committee on the Education and Utilization of the Engineer in this regaTd that is, industry, academia, and the professional societies need to join forces in promoting legislation wherever necessary to facilitate gifts of laboratory equipment to cot- leges of engineering. In the special case of Cricks and mortar," the federal government and industry should be prepared to match those funds raised for this purpose by state governments or from philanthropic sources. In addition, universities and the government should change their cost accounting practices to repect the faster real depreciation of equipment in the rapidly changing IC3fietd. Depreciation should be over S-7 years in the case of equipment, and 15-80 years in the case of buildings. CROSS-DISCIPLINARY RESEARCH The inclusion of research on IC3 systems within the purview of a single panel suggests the heavily interdisciplinary nature of this field. Like other areas of engineering research that are currently acquiring great economic and technological importance, work in lC3 systems cuts across traditional disciplinary boundaries. Yet the requirement for cross-disciplinary approaches to research and teaching runs counter to the established structures and practices of most university engineering departments, which have long empha- sized specialization. By the same token, cros~disciplinary research is not easily encompassed within the traditional academic depart- ment structure or the reward system for university faculty. These

210 DIRECTIONS IN ENGINEERING RESEARCH problems impede the needed transition to new modes of research and practice in the nation's schools and industries. Interdepartmental laboratories are a very useful organiza- tional mechanism within universities for dealing with research problems that span several departments or disciplines (e.g., VESI, robotics, manufacturing). One key to their success is the use of directors who are able to devote a substantial part of their time to management. Another great advantage is the presence of research staff members. The attractiveness of university life to these research staff members has been greatly reduced by many of the problems described earlier, in particular those of salaries and equipment. In addition, the status of these personnel within the university community is a matter for concern. Unlike the faculty members, they do not vote in university councils; nor are they a part of the academic policymaking process. Thus, they are to a great extent isolated from university life. As a result of these problems, they have recently been leaving universities in great numbers. Recommendation Universities must evaluate both their organizational and reward structures to permit the cross-disciplinary approach to Nourish, in research as well as in teaching. In addition, university admin- istrators must improve the salary structure for interdepartmental laboratory research staff and devise other mechanisms for inte- grating them into university affairs and otherwise improving their overall morale. HUMAN RESOURCES: ADEQUACY OF NEW TALENT THE B.S. AND M.S. Students entering the disciplines associated with IC3 include many of the very best students attending universities. Enrollment limitations in force at many institutions have raised the high- school grade point averages and Scholastic Aptitude Test scores of majors in electrical engineering and computer science to the high- est levels in memory (Horgan, 19843. The number of graduates in these disciplines at the B.S. and M.S. levels is at an all-time high and apparently still growing, despite signs of a downturn in engineering enrollments generally. There are now some reports of

IC3 SYSTEMS RESEARCH 211 declining demand for engineers even for those in the information and computer fields, in which growth has been phenomenal for a number of years (e.g., Inside RED, 1985; Office of Technology As- sessment, 1985~. However, the pane! believes that although there may be fluctuations, demand for these graduates will continue to grow for the next 5-10 years. Model-based projections tend to confirm this expectation (see, for example, Vanski, 1984~. THE PH.D. As we discussed earlier, a major reason for the current faculty shortage in this field is the shortage of new Ph.D.s. Although it has increased slightly in the past 2 years, the number of engineering doctorates is not substantially greater than it was in the late 1960s, and is, in fact, considerably below the level of the early 1970s. The number of Ph.D.s earned in computer science has remained level, at roughly 20~250 per year over the past decade, whereas the number in computer-relatecI areas of electrical engineering, for example, has been no higher than that. The quality of these doctorates is high; but their number is clearly insufficient- especially in ~hot" areas such as artificial in- telligence, CAD, robotics, VEST, computer architecture, graphics, and computer systems. The need for doctorates in these fields will not abate in the next decade. As noted earlier, computers and electronics are permeating all aspects of life and work; ~C3 Will continue to be research-intensive. In addition, whether the supply of B.S. ant] M.S. graduates comes into balance with demand or not, many more Ph.D.s will be needed to staff university faculties. The number of Ph.D.s could be increased greatly if more women sought doctorates in engineering. More to the point, the overall quality of engineers and engineering education could be raised if more women participated, bringing a new source of highly talented people into practice and teaching. {C3 does not seem to attract a large share of the women who do enter engineering, how- ever, and most of those who are in ~C3 are involved in software. The difficulty seems to be traceable to the early grades, in which a difference can be seen between boys and girls in the relative appeal of mathematics, laboratory exercises, and even the use of computers. It IS difficult to say what could be done to entice more women into advanced study in iC3. Women involved in research

212 DIRECTIONS IN ENGINEERING RESEARCH in both academia and industry can help by actively communicat- ing its excitement to their qualified female students and younger colleagues and by encouraging them to follow this path. The pane] notes that, among Ph.D. students currently study- ing at American universities, the proportion of foreign-born stu- dents on temporary visas to American-born students has risen sharply in recent years, to more than 40 percent (National Re- search Council, 1985~. It must be said, however, that these foreign-born students have provided many of the new young fac- ulty members who are in such short supply; some 25 percent of all junior faculty in engineering are reported to have taken the B.S. at foreign schools (Office of Technology Assessment, 1985~. Foreign-born graduates are also extremely valuable in U.S. industry a point that is often overlooked. It is frequently found that a large proportion of the engineering employees in the most advanced areas of R&D are foreign born. Training them in this country is thus a goof! investment. This access to some of the rest of the worId's best talent gives the United States an edge in international competition an edge the Japanese, for example, do not have. To discourage these people from staying (or, even more so, from coming) would reverse that advantage. The long-term health of lC3 in the United States requires a substantial increase in the number of Ph.D.s who can stay in the country to enter academia and industry. Attitudes toward doctoral study must change. The leading B.S. and M.S. graduates must be able to weigh the advantages of a Ph.D. against the alternatives and decide that it is worthwhile to pursue a doctorate. Recommendation More students must be induced to pursue the Ph.D. To that end, the pane! recommends that more substantial fellowships foe offered to American doctoral candidates, with a stipend equiva- lent to one-half the starting salary of an entry-level B.S. engi- neer in industry. In particular, the pane! commends initiatives by Hewlett-Packard and the American Electronic Association to award fellowships containing loans that are forgiven if the recipient remains in academia as a professor. Some quid pro quo in these fellowships might be useful; that is, requiring periods of work, reporting, or some other form of accountability in order to build a sense of responsibility in the recipients.

IC3 SYSTEMS RESEARCH References 213 Farber, D. Information Systems Engineering Perspectives. Paper presented at a National Science Foundation Workshop on Opportunities for En- gineering Research Focused on Emerging Engineering Systems, 15 July 1985. Haddad, J. A. Key issues in U.S. engineering education. The Bridge 13~2~:11- 16, 1983. Horgan, J. Technology '84 education. IEEE Spectrum 21:94-96, 1984. Institute of Electrical and Electronic Engineers, Inc. An IEEE Opinion on Research Needs in Information and Computing Technology. Report of an IEEE Task Force to the Engineering Research Board Panel on Information, Communication, Computation, and Control Systems Research, February 1985. Inside ROD, 14~6) :XXX, 1985. Editorial. National Research Council. Engineering Education and Practice in the United States: Foundations of Our Tcchno-Economic Future. Washington, DC: National Academy Press, 1985. National Science Foundation. Scicnec Indicators: 1982. Washington, DC: National Science Foundation, 1983. Office of Technology Assessment. Information Technology ROD: Critical Agenda and Induce (OTA-CIT-268~. Washington, DC: Office of Technology As- sessment, February 1985. Vanski, J. Projected labor market balance in engineering and computer specialty occupations: 1982-1987. In: Labor Market Conditions for Engi- necra: Is There a Shortage? Proceedings of a Symposium. Washington, DC: National Academy Press, 1984.

214 DIRECTIONS IN ENGINEERING RESEARCH Appendix Responses to the Engmeermg Research Board's Request for Assistance Tom Universities, Professional Societies, and Federal Agencies and Laboratories Requests for assistance were sent by the Engineering Research Board to a number of universities, recipients of Presidential Young Investigator awards, professional societies, and federal agencies and laboratories in order to obtain a broader view of engineering research opportunities, research policy needs, and the health of the research community. Some of the responses included comments on engineering research aspects of lC3 systems research; these were reviewed by this panel to aid in its decision-making process. The pane! found the responses to be most helpful and wishes that it were possible to individually thank all those who took the time to make their views known. Because that is not practical, we hope nevertheless that this small acknowledgment might convey our gratitude. Responses on aspects of IC3 systems research were received from individuals representing 53 different organizations, listed in Table A-1: 29 universities (including 11 represented by recipients of NSF Presidential Young Investigator Awards), 9 professional organizations, and 15 federal agencies or laboratories. Some com- ments coverer! specific aspects of the panel's scope of activities whereas others provided input on a variety of subjects. Although most of the responses addressed priority research needs, several respondents did reflect on policy issues. Many of the research needs and issues of policy and health addressed by the respondents were similar to those noted by pane! members. The broadened perspective provided by the responses to the survey was most beneficial in the panel's deliberations.

IC3 SYSTEMS RESEARCH TABLE A-1 Organizations Responding to Information Requests Relevant to Information, Communication, Computation, and Control Systems Research 215 UNIVERSITIES California Institute of Technology Carnegie-Mellon University Clarkson University Lehigh University Massachusetts Institute of Technology North Carolina State University Oregon State University Oregon Graduate Center Princeton University Rensselaer Polytechnic Institute Texas A&M University University of Arizona University of California, Berkeley University of California, Davis University of California, Los Angeles University of Colorado University of Georgia University of Hawaii University of Illinois—Urbana/ Champaign University of Iowa University of Kansas University of Maryland University of Michigan University of Oklahoma University of Pennsylvania University of Rochester University of Texas at Austin University of Utah Washington University PROFESSIONAL ORGANIZATIONS Association for Computing Machinery American Institute of Aeronautics and Astronautics American Institute of Chemical Engineers American Society of Civil Engineers American Society of Mechanical Engineers Industrial Research Institute Institute of Electrical and Electronic Engineers, Inc. Institute of Industrial Engineers Society of Engineering Science, Inc. AGENCIES AND LABORATORIES Air Force Institute of Technology Air Force Office of Scientific Research Argonne National Laboratory Army Research Office Lawrence Livermore National Laboratory NASA Ames Research Center NASA Goddard Space Flight Center NASA Jet Propulsion Laboratory NASA Lewis Research Center NASA Langley Research Center National Center for Atmospheric Research Nava! Research Laboratory Office of Naval Research Oak Ridge National Laboratory Sandia National Laboratory .

Next: 6. Manufacturing Systems Research in the United States: An Overview »
Directions in Engineering Research: An Assessment of Opportunities and Needs Get This Book
×
Buy Paperback | $75.00
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

Surveying the dynamic field of engineering research, Directions in Engineering Research first presents an overview of the status of engineering research today. It then examines research and needs in a variety of areas: bioengineering; construction and structural design; energy, mineralogy, and the environment; information science and computers; manufacturing; materials; and transportation.

Specific areas of current research opportunity are discussed in detail, including complex system software, advanced engineered materials, manufacturing systems integration, bioreactors, construction robotics, biomedical engineering, hazardous material control, computer-aided design, and manufacturing modeling and simulation.

The authors' recommendations call for funding stability for engineering research programs; modern equipment and facilities; adequate coordination between researchers; increased support for high-risk, high-return, single-investor projects; recruiting of new talent and fostering of multidisciplinary research; and enhanced industry support. Innovative ways to improve the transfer of discoveries from the laboratory to the factory are also presented.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!