II ●

Robot Technology
 11 ●

Robot Technology

Roots of Robotics Technology

The paper by Albus (see app. B, item 2) In the view of some computer science re-
surveys the state of robotics technology. searchers, robotics as a technology that will
Robots have a dual technological ancestry have significant social impact is still in its in-
that has an important effect on discussions fancy. They estimate that, given sufficient
about what they are, what they can do, and research support, they could produce a flexi-
how they are likely to develop. The two an- ble, intelligent robot for the market within
cestral lines are: 1) industrial engineering this decade. A robot of this type will be able
automation technology, a discipline that to move freely about an unstructured envi-
stretches historically over a century; and 2) ronment, and perform a wide variety of tasks
computer science and artificial intelligence on command with minimal reprogramming
technology that is only a few decades old. time.
Ideas about the nature of robots differ
according to the importance given to these
two technological roots. This view stresses the need for continuing
Most modern industrial robots are exten- basic research in computer science related to
sions of automated assembly-line technol- robotics, particularly in “artificial intelli-
ogy. This form of automation has not histori- gence. ” Robots are seen as “stand-alone,”
cally depended on computers, although mi- reprogramable devices, capable of perform-
croelectronics provides a powerful new tool ing many tasks other than large-scale assem-
for extending its capabilities. In this view bly line applications, for example, small-
modern industrial robots are closely related scale batch manufacturing, mining, or equip-
to numerically controlled machine tools. ment repair.

From such a perspective, robotics is

already approaching the state of a mature
Which of these views is most pertinent in
technology. Over the next decade, the most
terms of current policy issues will depend, in
important impacts of robotics on the econ-
part, on whether such an “intelligent” robot
omy and work force cannot be considered
would be economically feasible in the near
separately from the impacts of industrial
future and whether it would meet a signifi-
automation in general.
cant need in the industrial sector. It seems
On the other hand, modern computer tech- likely, in fact, that both types of robotics
nology may provide future robots with new technology will eventually become impor-
“intelligent” capabilities such as visual and tant, but that their economic and social im-
tactile perception, mobility, or understand- pacts will differ to the extent that they are
ing instructions given in a high-level, natural used for different purposes in different en-
language, such as “Assemble that pump!” vironments. Furthermore, the time scale for
The commercial availability of such capabil- widespread adoption will be significantly
ities may be one or two decades away. later for the “intelligent” machines.

7
8 ● Exploratory Workshop on the Social Impacts of Robot/es

Definition of Robots
It is difficult to establish a usable, general- variable programed motions for the perform-
ly agreed on definition of a robot. Experts ance of a variety of tasks.
use different approaches to defining the This definition seems to describe the current
term. The problem of definition is further state of the technology and is generally ac-
compounded for the public by images shaped cepted by U.S. industry.
by science fiction movies that bear no resem-
blance to robots currently on the market. Industrial robots have three principal com-
ponents:
At the same time, it is important to have
some common understanding of the term in 1. one or more arms, usually situated on a
order to define the state of the art, to project fixed base, that can move in several
future capabilities, and to compare efforts directions;
between countries. Depending on the defini- 2. a manipulator, the business end of the
tion used, for example, estimates of the num- robot, is the “hand” that holds the tool
ber of robots installed in Japan vary from or the part to be worked; and
3,000 to over 47,000 (5). This variation stems 3. a controller that gives detailed move-
in part from the difficulty of distinguishing ment instructions.
simple robots from the closely related “hard Computer scientists add to this list a few
automation”* technologies for transferring capabilities that are not generally commer-
material. cially available today, but that might be part
The Robot Institute of America, a trade of a general purpose robot of the future (6).
association of robot manufacturers and They include the following:
users, defines robots as follows: 4. locomotion some means of moving
A robot is a reprogramable multifunctional around in a specified environment;
manipulator designed to move material, 5. perception, the ability to sense by sight,
parts, tools, or specialized devices, through touch, or some other means, its environ-
ment, and to understand it in terms of a
task—e.g., the ability to recognize an
*The term "hard automation” refers to traditional custom obstruction or find a designated object
engineered automated lines. Although they may contain in an arbitrary location; and
some standard components, they are built to accomplish one
specific set of tasks and often must be completely torn down 6. heuristic problem-solving, the ability to
and rebuilt when the manufacturing process or product plan and direct its actions to achieve
design changes. higher order goals.

Technological Context of Robots

The principal technological context of ro- devices that performed tasks similar to
botics is the field of industrial automation. those done by modern industrial robots have
Most experts on industrial automation state existed for centuries. The principal dif-
that robots are only one component of a ference is that, whereas so-called “hard
large collection of related devices and tech- automation” is custom designed to a par-
niques that form the technological base of in- ticular task, robots are standardized, but
dustrial automation (7). This view was ex- flexible and programmable units that can be
pressed both at the workshop and in discus- installed in different environments with
sions of experts with OTA staff. Mechanical much less customization. (Some adaptation
 Ch. II—Robot Technology ● 9

is still often required). Clearly, there is a materials from station to station, oper-
tradeoff between the efficiency of hard auto- ate tools such as welders and spray
mation and the flexibility of robots. painters, and perform some assembly
tasks; and
Since machinery will be integrated with ● computerized information systems that
the total design of a factory it may not be keep track of inventory, trace the flow of
useful to distinguish robotics as an inde- material through the plant, diagnose
pendent technology. A fully automated fac- problems, and even correct them when
tory of the future might include the follow- possible.
ing components:
All of the above technologies are currently
● a computer-aided-design (CAD) system
under development and being used in some
that provides a tool for engineers to de-
form. They will likely evolve into compo-
velop new products on a computer using nents of a fully automated flexible manufac-
an electronic display screen. The data
turing facility.
base generated by the computer during
the design phase is then used by other Thus, there appear to be two parallel tech-
computerized parts of the factory; nological tracks along which industrial ro-
● numerically controlled machine tools bots are likely to develop: 1) stand-alone
and other automated devices that fabri- standardized units that will have varying
cate components of the product, trans- uses in many different environments; and 2)
port, and assemble them following robotics technology that is integrated into
instructions generated by the CAD complete factories that will, themselves, be
system; flexible. Any assessment of the impacts of
● robots, also operating under computer robotics would need to consider both types.
generated instructions, that transfer

The Robot Market

The current structure of the industrial ro- facturing such as mining or equipment
bot market—producers, users, and inves- repair.
tors–is discussed in detail in the back-
ground paper by Lustgarten (app. B, item 4). Domestic robot manufacturers appear to
fall into four grous:
The principal uses of robots today are spot
welding, spray painting, and a variety of so- 1. Traditional machine tool manufactur-
called “pick and place” operations that in- ers such as Cincinnati-Milacron that
volve simply picking up an object and put- have developed a robot product line.
ting it with a specific orientation in a pre- 2. Established firms such as Unimation
determined spot. that have specialized in industrial ro-
The automobile industry is the largest bots.
user of industrial robots, in terms of the 3. Large manufacturing firms, such as
value of equipment installed, and probably General Electric and, in particular, elec-
will continue to be over the next decade. tronic computing equipment manufac-
Other major current and potential future turers such as Texas Instruments, that
users are summarized in the Lustgarten pa- plan to be major users of robots and that
per. Once again, these estimates consider the have decided to build their own. These
industrial robot as an extension of manufac- firms may choose either to retain the
turing equipment. They do not consider technology for their own use or to mar-
possible new applications outside of manu- ket their robots externally.
10 . Exploratory Workshop on the Social Impacts of Robotics

4. Small entrepreneurial firms that devel- innovative new types of robots will come
op new, innovative robots. This type of from the entrepreneurs, while the large firms
firm has been important in many sec- will have the capital and capacity to produce
tors of the information industry, and and market large quantities of heavy equip-
could well play an important role in ment. Also significant in this regard is the
robotics. trend, common with most high technology
firms, toward acquisition of small, innova-
The relative importance in the market-
tive firms by larger industrial firms seeking
place of these different types of firms will de-
either to diversify or to integrate their tradi-
pend on and, in turn, influence the evolution
tional product lines with new technologies.
of robotics technology. The history of the
microelectronics market suggests that many

Technology and Market Issues

A number of issues concerning the robot Government use. —Are there particular-
industry were identified in this project: ly important applications of robots in
the Federal Government that should be
● Industrial organization.—What types explored and developed? Experts at the
of firms will play the most significant workshop mentioned in particular de-
role in the production of robots and in fense applications and uses of robots
innovation? Will robot use and produc- for space exploration and oceanograph-
tion be concentrated in a few large com- ic work.
panies? Will a variety of robotics prod- Definition. –The question of defining
ucts be available for many applications robotics and their context, while not a
by diverse types and sizes of users? policy issue per se, is an important
What will be the effects on the financial problem if any Federal action is con-
health of different types of potential templated to encourage their use or de-
producers and users? velop any R&D program. How the tech-
● Research and development (R&D).– nology is defined may well determine
Should R&D stress applications or the type of industry that will be helped
should it focus on fundamental work by the programs, and influence the
aiming at significant new breakthrough structure of the U.S. robotics industry.
in the state of the art? What role should Standards. –Should the Government
the Federal Government play in fund- encourage the establishment of tech-
ing this research via agencies such as nical standards for robotics devices and
the National Science Foundation? components? Should standards be set
What type of work should be pursued in for interfacing between robots and
Government research labs such as the other automation and information tech-
National Bureau of Standards, and at nology? Would standards encourage
what level should it be funded? What the development of the robot industry
additional policies, if any, would be re- and the diffusion of the technology, or
quired to stimulate R&D in the private would they prematurely freeze the state
sector? of the art? -