Robotics: The big picture

Robots are hot these days. It's a far cry from the '80s, when automakers and other manufacturing giants saw their expensive experiments with robots fail, and manufacturing executives declared the industry dead. Robots were unreliable, they said. They were a .solution looking for a problem," an expensive replacement for cheap labor. North American sales were only $500 million in 1990.

Now automakers looking for flexible automation to make products cheaper, faster, and better have come back to robots. Industry boosters say robots have ceased to be a luxury and have become a tool to offset labor costs, absenteeism, shortages of trained workers, and human error. The North American robotics companies that cut staff in the '80s have been hiring and building new plants. By 1997, North American sales exceeded $1 billion, the best year ever, according to the trade group Robotic Industries Assn. (RIA; Ann Arbor, MI). In the first quarter of this year, 4732 robots valued at almost $363 million were sold, up 67% in numbers of units and 24% in value from the first quarter of 1998.

And the numbers will keep going up, the experts say. The United Nations Economic Commission for Europe, which reports on robotics around the world, predicts worldwide investment in industrial robots will rise 40% in 2001 over 1997, with 870,000 robot units at work in 2001, Western Europe will increase its robot stock 50%, the UN predicts. In the ITS, robot sales will be tip 50%
UN analysts say that the balance of automation seems to be swinging away from Japan to Europe and North America, as Japan, long a major user of robots, cuts back investments in the wake of the Asian economic crisis.

About half the robots sold still go to automotive applications like spot welding, those "dirty, dumb, and dangerous" jobs that auto plants need done. Material handling applications for a range of industries is also fueling the spurt in robot use, according to Donald A. Vincent, executive vice president of RIA.

Six factors are driving the robot boom, according to Dr. Hadi Akeel, senior vice president and chief engineer at Fanuc Robotics (Rochester Hills, MI) and one of RIA's industry observers. They are: the growing power of the microprocessor and the reduction of digital-processing costs; the appearance of low-cost and easy-to-use sensors; ability to test software for reliability; the multitasking operating environment and powerful off-the-shelf drivers typical of the PC; mature software for graphic processing, speech recognition, digital communications, and real-time control for interface devices; and developments in micro-machining.

Dr. Walter E. Red of Brigham Young University, in a report to SME's Robotics International association on technology trends this year, also stresses open architecture and plug-ins. Digital control technologies, he says, are driving highspeed networks that permit distributed architectures for motion planning and control and for digital drives. Robots are moving faster and optimizing the execution of joint and curvilinear moves and their transitions. Calibration techniques that filter path and position data within the controller allow adjustments for inaccuracies in robot moves to real-world positions. Greater robot accuracy means closer correlation between robot performance and simulation studies.

Motion systems are becoming more tightly linked to the process. Multiprocessor and multitasking architectures have moved robot controllers "from a world of autonomous function to a networked world where information can be collected and dispensed," says Red. PC interfaces now provide software modules called "robot servers" that give access to the robot's internal parameters.

Off-line programming of robotic work cells is beginning to fulfill its promise, thanks to static and dynamic accuracy compensation algorithms. As the PC industry develops fast processors like the Pentium III 550 MHz and software developers produce packages for the PC, simulation will become available to everyone.

RIA's Vincent thinks the robots at work in US automakers' plants are just the first wave of users. "Thousands of companies that could benefit from robots, particularly small and medium-sized companies, haven't purchased even one robot yet," he says. He believes that those manufacturers will soon get on the bandwagon, "as the cost of robots continues to fall and companies become more aware of how much they can improve productivity, and quality." (An article in this issue, "Robotic Welding Downsizes," takes a look at vendors' efforts to reach this group of manufacturers.)

Despite these technological strides, the experts at SME's Robotics International warn against overoptimism. In this year's report, Elaine Hinman-Sweeney, principal software engineer at Raytheon's OBCS/Robotics, says perceptions of robot effectiveness depend on preconceptions of how robots should be used. She points out that no engineer or manager would seriously produce a process plan for machining a high-performance engine on sophisticated CNC machining centers alongside a plan for drilling, milling, boring, and grinding that engine with hand tools if the machine tool malfunctions. No engineer or manager at a stamping plant would draw up plans to make stamped metal products with mallets and wooden forms in case the plant's presses or dies malfunctioned.