Click for next page ( 14


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 13
Toward the "Third Generation" of Intelligent Robots Although the United States played a leadership role in the development of robotics technology, U.S. manufacturers by and large failed to capitalize on this advanced form of automation. Unable to interest domestic manufacturers in their robots, pioneering fops such as Unimation licensed their technology to Japanese companies. That technology was applied broadly, beginning in the automobile plants. Robot lines were installed in parallel with manual production lines in Nissan's Okama plant, and Nissan and Kawasaki worked together to improve robotics technology. Accompanying these improvements was a vigorous emphasis on quality control. U.S. manufacturers have lagged, and continue to lag, behind their Japanese counterparts in the adoption of robotics, in part because of their investment crite- ria. Whereas large U.S. firms insist on a return on investment of 15 percent, Japanese firms accept a 3 percent return on investment and often achieve greater profitability. In Japan structural relations between developers and users seem to matter more than individual company decisions. This tends to promote long-term planning and to foster a concomitant willingness to make investments that will not pay off for some time. Toshiba, for example, derives considerable support from the electric power industry and the Japanese government for much of Me work it is doing in nuclear applications. A shortage of skilled labor and a union system (industry- as well as company- wide3 that promotes labor-management cooperation has enabled Japanese compa- nies to maintain steady employment in the manufacturing sector while introducing substantial numbers of robots. Studies by MITI, Japanese labor unions, and others suggested a minimal impact of robots on employment, countering the International 13

OCR for page 13
14 0% 50% 100% Metal Products Industrial Machinery Electrical Machinery Automobiles Precision Equipment Synthetic Plastics Processing Other Manufacturing ~2 ',,,2,. IDD Cost Increase of 1 Million to 2 Million Yen em Cost Increase of Less Than 500,000 Yen Virtually No Cost Savings Cost Savings of Less Than 500,000 Yen Cost Savings of 500,000 to 1 Million Yen Cost Savings of 1 Million ED 2 Million Yen Cost Savings of 2 Million lad 3 Million Yen Cost Savings of Over 3 Million Yen FIGI5RE3 Cost savings do - yip ~:J~R~ Metallurgy Federation's 1979 findings that anticipated a negative impact. Japanese labor and management reached a consensus in 1980, and discussions about the negative impacts of robotics effectively ended. Japanese workers, being relatively secure in their jobs, more readily accept and appreciate the contribution that robots make to their firms' competitiveness. Competition is fierce among Japanese companies, and labor savings from the introduction of robots have significantly decreased the cost of production (see Figure 3~. Robots have reduced the number of persons required per shift and increased the number of shifts possible (see Figure 4~. In some industries cost sav- ings from robotization have been significant. Japanese government and industry also anticipated the need to prepare for the introduction and maintenance of robots. Recognizing the need for more workers in planning and maintenance, and for a shift from mechanical to electrical and electronic engineers, many Japanese feds adjusted their recruiting policies and placed greater emphasis on employee training. The fact that high quality is best achieved by a balance of technology and people was corroborated by a context- oriented (work organization, human resource, and management strategy) study by the Massachusetts Institute of Technology of 80 auto plants in 15 U.S. cities. The study concluded that technology was a weak predictor of productivity compared to quality and management. By 1984 the motivation in Japan for using robots had shifted from labor cost savings and response to an aging society toward flexibility for small-lot, multi- product lines and productivity improvement. Manufacturing companies began to perceive a need to supplement simple special-purpose robots with intelligent

OCR for page 13
15 o 2 - Metal Products 0.9 ~ (49) 0.9 ~ (45) industrial Machinery Electrical Machinery Automobiles Precision Equipment and Plastics Processing Other Manufach'~ng Total 1.31 (68) 1 .1(42) i.O(41 ) . o.s1~32~ 1 11 (Number ofRespondir~g ' Companies: 277) Persons per Robot per Shift FIGURE 4 Average production worlcer reduction per robot per shift (per canpany). SOURCE: Japan lddustnal Robot Association. robots with memory and decision-making capabilities. The automobile industry, for example, came to consider increased automation of the final assembly process essential to improved vehicle performance and product competitiveness. (The shift in applications of robots in the automobile industry is illustrated in Figure 5.) Intelligent robots that corresponded to elements of the assembly process were needed. To assemble elusive matching parts in confined spaces while avoiding complex obstacles, such robots would have to be equipped with higher-level con- trol technologies, possess visual compensation and force control, and be capable of multirobot cooperation. Improved robot arms, tools, and a more sophisticated man- machine interface would be required. The needed R&D has been undertaken in many settings. Toshiba, a large Japanese integrated electronics manufacturer, employs 15,000 of its 70,000 employees in corporate R&D and engineering laboratories, many of which are engaged in robotics-related hardware and software development. Fifteen- to 20- year projects that would not be feasible in the United States, such as the develop- ment of certain robotics applications in the nuclear power plant industry, are undertaken by companies like Toshiba with support from the electric power indus- try and the government. Although strong formal ties generally do not exist between industry and univer- sities in Japan, informal ties are very strong. University professors typically sub- mit research themes to the Ministry of Education in hopes of receiving small amounts of funding. Industrial funding through sponsored research and donations would be helpful, but Japanese universities have few full-time research staff mem- bers with master's or doctoral degrees and little experience coordinating large

OCR for page 13
16 6,Oo0 5,000 ,000 2,000 1 ,000 _ ,j, Plastic Molding Casting _ J Painting | Arc Welding | ~ '' _ Welding ~ ~ Sub ;titution for Dedicated Machines Assembly 1988 Future (10 years) 1978 YEAR FIGURE 5 Introducuon of robots. SOURCE: Toyota hIotor Corporanon. lligent Robots Substitution for Manual Operation (includes intelligent robom) R&D projects with industry. Trial organizations, such as the University of Tokyo's Research Center for Advanced Science and Technology, which was orga- nized as a mechanism for raising industrial capital, are still in the early stages of development In general, industry is driving robotics R&D in Japan, with considerable sup- port from the government. One example of Japanese government-industry cooper- ation is the National Research and Development Program (popularly known as the Large-Scale Project). Initiated by the Agency of Industrial Science and Technology in 1966, the program undertakes risky large-scale R&D on a commis- sion basis under government leadership. Projects must address important and urgent needs and potentially benefit Japanese industry or society. A MITI adviso- ry council plays a role in the selection of projects. Implementation is divided between national laboratories (which conduct basic research) and the private sec- tor (which pursues development). Government-commissioned user associations are established in the private sector. Equipment bought by industry with govern- ment funds becomes a national asset. Project duration is typically 6 to 7 years. Ten projects are currently under way, including the Advanced Robot Technology project, which is directed at developing technologies that will provide the maneu- verability and general-purpose adaptability needed to commercialize robots for

OCR for page 13
17 inspection, maintenance, rescue, and other broadly diversified jobs, and the Automated Sewing System project, which is aimed at developing an integrated production system for small-lot production, materials handling, and sewing tech- niques to reduce production time for small- to medium-sized textile companies. Three robotics projects undertaken in the United States illustrate a range of objectives, funding, and management approaches. The remote technology used for surveying and recovery at the crippled Three Mile Island nuclear reactor was developed in an applied program with specific objectives and was funded princi- pally by the federal government. The Autonomous Land Vehicle was a govern- ment initiative that developed advanced, fundamental robot vision, planning, and navigation technology without an immediate application focus. Current NASA programs aimed at developing robotics applications for space exploration exempli- fy a healthy balance of basic and applied activity, focused by objective on mission scenarios, and with the latitude to build a base of progressive, future-oriented knowledge. The National Science Foundation (NSF) has a loose strategy for funding basic research to support development of manufacturing, construction, and intelligent robotics systems and provides seed money for the dissemination of results. Needs are articulated through requests for proposals, workshops, and suggestions from study groups and other government agencies. The National Institute of Standards and Technology (NISI) funds mainly its own activities while making small grants to universities. Needs are articulated internally by an oversight committee and externally as requests from other govern- ment agencies. NIST's R&D has had a major impact on control systems strategies. The U.S. DeparOnent of Defense's (DOD) Joint Service Robotics Panel shares information but does not coordinate activities. Each service looks at robotics in terms of its respective mission. DOD, He largest source of funds for robotics, stimulates research through the Office of Naval Research, the Air Force Office of Scientific Research, the equivalent office of He Amity, university research initia- tives, and the NSF. If technical requirements of DOD's service requests call for developing enabling technologies, the work is routed to the Defense Advanced Research Projects Agency. In addition, the departments of Energy (DOE), Health and Human Services (HHS), and Interior (DOI) are all engaged in robotics research. DOE research focuses on nuclear and hazardous waste, HHS on applications for the handicapped and in prosthetics, and DOI on mining and undersea applications. In all three agencies, management identifies needs, which are evaluated internally. Most industry-university cooperation in the United States involves industry funding of university-based projects. Resources and research needs may not coin- cide in university settings. Institutions such as Carnegie-Mellon University, which has a strong robotics program, must compete for resources. Unmanned excavation, for example, though perceived as a need at Carnegie-Mellon 4 years ago, was not funded. Eventual funding by He NSF was insufficient to support the initiative.

OCR for page 13
18 The United States hosts mainly basic research on dynamic control, machine perception, artificial intelligence, and neural networks. Funding is largely from the federal government, with development carried out at open universities. Because the U.S. government funds comparatively little applied research in robotics, and only defense programs fund development in industry, transfer of the universities' basic research into products has been slow. In Japan robotics research is largely applied, and there is more cooperation among government, industry, and academe. Programs include undersea, construction, and health care robots. Japanese researchers benefit from attendance at those U.S. universities where basic research is carried out, but there is no mechanism by which U.S. com- panies can benefit from applied research conducted in Japan. These differences in how robotics technology is developed and commercialized in the United States and Japan-differences in the focus of research, sources of funding, and mecha- nisms for technology transfer-can be viewed as a platform for cooperation. The United States and Japan do not have a history of successful cooperation in the development of robotics technology. U.S. technology has gone to Japan via licensing. Japanese companies have insisted on manufacturing licenses while refusing to distribute U.S.-made products. Many U.S. companies believe that Japanese robotics companies are reluctant to license their technology to U.S. com- panies for manufacture. Past proposals for collaborative efforts have generated skepticism among U.S. partners, even though they could theoretically learn from Japanese companies about how to manage cooperation. Future efforts must, there- fore, demonstrate clear benefits to both sides. Because of these factors, many believe that cooperative work can best begin in noncompetitive areas where there are no likely immediate commercial impacts but where the goals are clearly defined and attainable within a few years. Unless the two major players in this arena learn to cooperate, the world is not likely to benefit any time soon from the capabilities of disaster recovery robots and personal robots or the care of the elderly and infirm.