The Robotic Industries Association, the leading trade group for the robotics industry, defines a robot as follows: it is a "reprogrammable, multifunctional manipulator designed to move material, parts, tools or specialized devices through variable programmed motions for the performance of a variety of tasks." This definition has become generally accepted in the United States and other Western countries. The most common form of industrial robot is made up of a single automated arm that resembles a construction crane.
The word "robot" was coined by Czech playwright Karel Capek (1890-1938) in his 1921 play R.U.R. (Rossum's Universal Robots). Robot is spelled robota in Czech and means forced labor. The word found its way into English-language dictionaries by the mid-1920s. The word "robotics" was first used by science fiction writer Isaac Asimov (1920-92) in his 1942 story "Runaround," in which he wrote what became known as Asimov's Three Laws of Robotics: "1. A robot may not injure a human being, or, through inaction, allow a human being to come to harm. 2. A robot must obey the orders given it by human beings except where such orders would conflict with the First Law. 3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Law." Though fictional, these laws and Asimov's robot stories were influential to Joseph Engelberger, who is arguably the most important figure in the development of industrial robots. Though the word "robot" is relatively new, the concept is centuries old, and prior to the 1920s robot-like mechanisms were called automatons. In one of Noah Webster's earliest dictionaries, an automaton is defined as "A self-moving machine or one which moves by invisible springs."
In a number of respects, robots are like numerically controlled automated machine tools, such as an automated lathe, since they are both reprogrammable to produce a number of different objects. What distinguishes robots is their flexibility regarding both range of tasks and motion. In one typical manufacturing application, robots move parts in their various stages of completion from one automated machine tool to the next, the system of robots and machine tools making up a flexible manufacturing workcell. Robots are classified as soft automata, whereas automated machine tools are classified as hard automata. The Japanese Industrial Robot Association also classifies manually operated manipulators and nonreprogrammable, single-function manipulators as robots, and one must bear this in mind when comparing data on robot use between Japan and the United States.
Since robots are defined by their capacity to move objects or tools through space, key issues in robotic control are location and movement, referred to in the industry as kinematics and dynamics. The position of an object in a three-dimensional space can be defined relative to a fixed point with three parameters via the Cartesian coordinate system, indicating placement along x, y, and z axes. The orientation of an object requires three additional parameters, indicating rotation on these axes. These parameters are referred to as degrees of freedom. Together these six parameters and the movement among them make up the data of kinematic control equations. Robots carrying out simpler tasks may operate with fewer than six degrees of freedom, but robots may also operate with more than six, which is referred to as redundancy. Redundancy gives a robot greater mobility, enabling it to more readily work around obstructions and to choose among a set of joint positions to reach a given target in less time.
Two types of joints are commonly used in robots, the prismatic or sliding joint, resembling a slide rule, and the re volute joint, a hinge. The simplest type of robot to control is one made up of three sliding joints, each determining placement along a Cartesian axis. Robots made solely of revolute joints are more complex to control, in that the relation of joint position to control parameters is less direct. Other robots use both types of joints. Among these, a common type uses a large sliding joint for vertical placement of an arm made of revolute joints. The vertical rigidity and horizontal flexibility of such robots make them ideal for heavy assembly work (this configuration is referred to as SCARA for Selectively Compliant Arm for Robot Assembly). Robots may also be made of a system of arms each with restricted movement (i.e., with relatively few degrees of freedom) but which together can perform complex tasks. These are referred to as distributed robots. Such robots have the advantage of high speed and precision, but the disadvantage of restricted range of movement.
Robots are activated by hydraulic, pneumatic, and electrical power. Electric motors have become increasingly small with high power-to-weight ratios, enabling them to become the dominant means by which robots are powered. The hand of a robot is referred to in the industry as an end effector. End effectors may be specialized tools, such as spot welders or spray guns, or more general-purpose grippers. Common grippers include fingered and vacuum types.
One of the central elements of robotics control technology involves sensors. It is through sensors that a robotic system receives knowledge of its environment, to which subsequent actions of the robot can be adjusted. Sensors are used to enable a robot to adjust to variations in the position of objects to be picked up, to inspect objects, and to monitor proper operation. Among the most important types are visual, force and torque, speed and acceleration, tactile, and distance sensors. The majority of industrial robots use simple binary sensing, analogous to an on/off switch. This does not permit sophisticated feedback to the robot about how successfully an operation was performed. Lack of adequate feedback also often requires the use of guides and fixtures to constrain the motions of a robot through an operation, which implies substantial inflexibility in changing operations.
Robots may also be able to adjust to variations in object placement without the use of sensors. This is enabled by arm or end effector flexibility and is referred to as compliance. Robots with sensors may also make use of compliance.
Robots are programmed either by guiding or by off-line programming. Most industrial robots are programmed by the former method. This involves manually guiding a robot from point to point through the phases of an operation, with each point stored in the robotic control system. With off-line programming, the points of an operation are defined through computer commands. This is referred to as manipulator level off-line programming. An important area of research is the development of off-line programming that makes use of higher-level languages, in which robotic actions are defined by tasks or objectives.
Robots may be programmed to move through a specified continuous path instead of from point to point. Continuous path control is necessary for operations such as spray painting or arc welding a curved joint. Programming also requires that a robot be synchronized with the automated machine tools or other robots with which it is working. Thus robot control systems are generally interfaced with a more centralized control system.
Industrial robots perform both spot and electric arc welding. Welding guns are heavy and the speed of assembly lines requires precise movement, thus creating an ideal niche for robotics. Parts can be welded either through the movement of the robot or by keeping the robot relatively stationary and moving the part. The latter method has come into widespread use as it requires less expensive conveyors. The control system of the robot must synchronize the robot with the speed of the assembly line and with other robots working on the line. Control systems may also count the number of welds completed and derive productivity data.
Industrial robots also perform what are referred to as pick and place operations. Among the most common of these operations is loading and unloading pallets, used across a broad range of industries. This requires relatively complex programming, as the robot must sense how full a pallet is and adjust its placements or removals accordingly. Robots have been vital in pick and place operations in the casting of metals and plastics. In the die casting of metals, for instance, productivity using the same die-casting machinery has increased up to three times, the result of robots' greater speed, strength, and ability to withstand heat in parts removal operations. In 1992, CBW Automation Inc. of Colorado announced the development of the world's fastest parts-removal robot for plastics molding. Their robot moves through a four-foot stroke in under one-fifth of a second.
Assembly is one of the most demanding operations for industrial robots. A number of conditions must be met for robotic assembly to be viable, among them that the overall production system be highly coordinated and that the product be designed with robotic assembly in mind. The sophistication of the control system required implies a large initial capital outlay, which generally requires production of 100,000 to 1,000,000 units per year in order to be profitable. Robotic assembly has come to be used for production of printed circuit boards, electronic components and equipment, household appliances, and automotive subassemblies. As of 1985, assembly made up just over ten percent of all robotic applications.
Industrial robots are widely used in spray finishing operations, particularly in the automobile industry. One of the reasons these operations are cost-effective is that they minimize the need for environmental control to protect workers from fumes. Most robots are not precise enough to supplant machine tools in operations such as cutting and grinding. Robots are used, however, in machining operations such as the removal of metal burrs or template-guided drilling. Robots are also used for quality control inspection, to determine tightness of fit between two parts, for example. The use of robots in nonindustrial applications such as the cleaning of contaminated sites and the handling and analysis of hazardous materials represent important growth markets for robotics producers.
The first industrial robot was developed in the mid-1950s by Joseph Engelberger (1925), who has been referred to as the father of industrial robots. Engelberger also founded Unimation, Inc., which became the largest producer of industrial robots in the United States.
His early research involved touring Ford, Chrysler Corp., General Motors, and 20 other production plants. Engelberger observed that men performed the higher-paying jobs in which they lifted heavy objects with two hands simultaneously, while women performed tasks in which they used their hands asynchronously. Economic and technical considerations thus led Engelberger to focus on the development of a one-armed robot. Engelberger developed his first prototype in 1956, the design of which is very similar to Unimation robots produced decades later.
General Motors purchased a test model in 1959, though by 1964, Unimation had sold only 30 robots. It was not until the late 1960s that sales increased strongly and not until 1975 that the firm turned a profit. Together with number-two producer Cincinnati Milacron, Unimation accounted for 75 percent of the U.S. robotics market in 1980. Unimation Inc. became a wholly owned subsidiary of Westinghouse in 1982. By 1983, the firm had sales of $43 million.
Generally, there was much greater reluctance to adopt the use of robots in the United States than in places like Japan, which led the world in robot production and use. Among other apprehensions, U.S. companies balked at the heavy investments required and were sensitive to opposition from organized labor. Other times, when robots were ordered, they failed to deliver what manufacturers anticipated, due to unrealistic expectations of flawless operations and dramatic labor savings—ignoring the heavy maintenance robots required—and to the limitations of the machines themselves.
While the early 1980s saw promising growth, yielding some predictions that robotics would be a multibillion-dollar business by 1990, the U.S. robotics industry suffered a severe setback in the mid-1980s, largely the result of declining orders from the automobile industry. At the time, the auto industry still supplied over 70 percent of all U.S. robot orders. This resulted in a number of firms leaving the industry, including deep pocketed players such as General Electric and Westinghouse, and left many of the remaining firms to merge or be acquired. The value of new orders fell from their 1984 high of $480 million for 5,800 units to a 1987 low of just 3,800 units worth $300 million.
After staging a strong recovery in the late 1980s, the industry faced another setback during the recession of the early 1990s. Orders stagnated in 1990 at $510 million based on 5,000 units industry-wide, and slumped to 4,000 units at $410 million the following year. However, following the general economic recovery, business picked back up in 1992, and by 1993 the industry received record orders for some 6,800 robots valued at $630 million. Solid orders and shipments continued throughout the remainder of the 1990s, often at double-digit growth rates, and surpassed in 1997 the billion-dollar mark for the first time. That year, companies ordered 12,149 robots worth $1.1 billion. Though orders slowed slightly in 1998, they remained strong into 1999. As of 1999, more than 92,000 robots were in operation in U.S. industrial settings.
A corollary to the industry's robust economic performance in the 1990s was that the industry was now building and marketing its products differently. Whereas earlier robots were ambitiously designed to take on giant tasks, but couldn't necessarily do so with great precision and reliability, designers increasingly focused their robots on performing more manageable tasks with greater consistency. They also made robots easier to operate and maintain. As reliability and ease of use were—and are—some of the biggest concerns companies have about robots, this helped fuel demand and improved the robotics industry's image. Robot manufacturers were also more careful not to promise more than their devices could deliver, a common complaint lodged against them during the 1980s.
Lower prices also contributed to the sales surge. Whereas in 1984 the average robot cost an estimated $82,758 (net value of orders divided by the number of units), by early 1999 the average had fallen to $76,669, not accounting for inflation. After inflation is factored in, the real reduction in prices was more than 40 percent over the 15-year period.
The robotics industry has also diversified its customer base. While automotive-related manufacturing still accounted for about half of the U.S. market in 1999, inroads were being made in non-automotive materials handling, flexible manufacturing systems, and service-oriented uses. Some of the other major industry sectors purchasing robots include electronics manufacturing, food and beverage production, pharmaceutical manufacturing, and the aerospace industry.
Some of the largest robotics companies operating in the United States include ABB Flexible Automation, Adept Technology, Inc., and Fanuc Robotics. As of the late 1990s, Adept Technology was the only major U.S.-based manufacturer, while ABB was part of a Swiss-based conglomerate and Fanuc, the world's largest robotics company, was based in Japan. Fanuc also had a U.S. joint venture with General Electric called GE Fanuc Automation North America.
Meanwhile, Japan has continued to dominate the world robotics arena. Prior to 1978, the largest user of industrial robots in Japan was the automobile industry, after which the electric and electronics industries became most important. Its production of industrial robots quadrupled from the mid1980s to 1990, when it possessed over 40 percent of all industrial robots in use worldwide. By the late 1990s, annual Japanese orders for new robots still outpaced U.S. orders by a nearly four to one ratio. Unlike the United States, Japan makes extensive use of SCARA configuration robots, which offer substantial advantages in assembly work.
Recent research and development has addressed a number of aspects of robotics. Robotic hands have been developed which offer greater dexterity and flexibility. Most visual sensors in use were designed for television and home video, and do not process information quickly for optimal performance in many robotics applications. Consequently, solid-state vision sensors came into increased use, and developments were also made with fiber optics. The use of superconducting materials offered the possibility of substantial improvements in the electric motors that drive robotic arms. Attempts were made to develop lighter robotic arms and also to increase their rigidity. Standardization of software and hardware to facilitate the centralization of control systems was also an important area of development, as was miniaturization of parts to create smaller devices.
One example of a relatively successful service robot is a series called HelpMate by Helpmate Robotics Inc., headed by industry patriarch Joseph Engelberger. The HelpMate serves as an in-house courier in hospitals, delivering reports and lab samples to various destinations throughout the building. Each unit has the building's floor plan stored in memory, and includes sensors so it doesn't run into people or other objects. In 1999 HelpMate units had been installed in approximately 70 U.S. hospitals. Other service-related areas of robotics under development include surgical devices and in-home personal-care assistants for elderly and disabled persons.
[ David Kucera ]
Asimov, Isaac, and Karen Frenkel. Robots: Machines in Man's Image. New York: Harmony Books, 1985.
Loizos, Constance. "Domo Arigato, Mr. Roboto." Red Herring, May 1998. Available from www.herring.com .
"Robot Orders Rocket to New Record in First Quarter." Robot Times, summer 1999.
Roush, Matt. "Robot Industry Expands from Manufacturing." Crain's Detroit Business, 19 April 1999.
Vincent, Donald. "The Robotics Industry: Leading the Charge to a Productive 21st Century." Ann Arbor, MI: Robotic Industries Association, 1999. Available from www.robotics.org .
Winter, Drew. "The Droid Void Is Ending." Ward's Auto World, January 1994, 56.