``Up. Down. Clockwise. Counter.'' Rick Gosz is reading from a list of words flashing in front of him on a personal-computer screen.
Speaking into the microphone of his lightweight operator's headset, Mr. Gosz is saying ``hello'' to Command-1, a unique new PC-controlled, voice-activated robot, teaching it to recognize the sound of his speech as he runs through a series of commands.
In less than 10 minutes, Gosz, who has muscular dystrophy, is ready to go. This time, as he speaks, Command-1's robot arm instantly obeys his orders, swinging left, then down, then back, then up, it's ``hand'' reaching for and grabbing a pen.
``It's relatively easy to use and it picked up on my words rather quickly,'' he says with surprise.
A computer engineer himself, Gosz was tapped from the crowd at a recent show of robotic technology in Detroit to demonstrate how the $40,000 system from Prab Command Inc. in Chicago can be used by people with limited mobility who want to keep working.
Command-1 and a related system, Command-2, represent perhaps the most unusual use to date for robotic systems. Until recently, these electro-mechanical devices were mainly limited to industrial uses: welding car bodies or plucking parts from a conveyor belt, for example. But as the Robots 12 conference in Detroit demonstrated, the group known as ``service robots'' is moving into a variety of new fields.
``The market for service robots could eventually dwarf the industrial robot,'' says Walter Weisel, Prab's chairman.
There are some 600,000 people with varying degrees of physical handicap in the United States, and it typically costs the state, insurance companies, or their families $750,000 to support each one for life, from the time of their injury. More significant is the loss of ``the brain trust'' of the handicapped who not working.
Command-1 is just going into production, and Mr. Weisel expects to ship about 40 this year for limited field testing.
The Command-2, another new Prab robot, uses a mechanical arm to provide therapy to disabled patients. It recently completed five months of testing at the Rehabilitation Institute in Detroit, and will also be going on sale soon.
A robot built by another company, Odetics Inc. of Anaheim, Calif., ``walks'' on six ``legs,'' giving it the ability to step over obstacles. The company hopes to use the device to enter and perform repairs inside the most dangerous areas of nuclear power plants.
For several months, another new service robot has been going through tests at a Danbury, Conn., hospital, running errands, delivering lab samples, and taking over many of the basic duties that would otherwise occupy a nurse or orderly.
Dubbed HelpMate, the $40,000 system will also be introduced this year by Transitions Research Corporation, a company founded by the man many credit with giving birth to the modern field of robotics.
``If you look at industrial robots, they're attacking the jobs done by 25 percent of our work force, but 75 percent of our work force are involved in service, and many of those jobs are even more distasteful than the worst factory work ... like cleaning toilets,'' says Joseph Engelberger, the chairman of TRC and an industry pioneer for 30 years.
With his close-cropped hair, bright bow tie, and energetic, outgoing manner, Mr. Engelberger has been the industry's willing ambassador to the public. Today, he says, the idea of taking robots off the factory floor and putting them into the typical home is becoming less and less a matter of science fiction.
Until now, service robots have been little more than toys, such as the $4,495 HERO produced by Heath-Zenith Computers & Electronics, which can do little more than roll around the floor carrying a tray.
While thinking robots like R2D2 of ``Star Wars'' may still be light years away, TRC is completing the first phase of a multi-million-dollar study that could produce a truly functional robot butler by the early 1990s. Interestingly, the research is being funded by a seemingly unlikely consortium of six household goods manufacturers, including Johnson Wax, Maytag, and Electrolux.
The first phase of the study will be completed late this summer, Engelberger notes, optimistically adding that he expects the consortium immediately to begin funding the actual development process.
If all goes well, Engelberger expects to see the first working model in use by 1991. ``It will be a luxury product in the mid-1990s,'' he says, adding that as more and more robots get into production, their costs will come down dramatically.
These household ``butlers'' could handle a variety of duties: cooking, cleaning bathrooms, security, even washing the car. But Engelberger says they won't look much like the anthropomorphic R2D2 of ``Star Wars.'' They will still roll around on wheels, for one thing, and that will require that robot-equipped homes be designed to accommodate the machines, with smooth floors on one level.
Today, service robots account for less than $120 million a year in sales, but Engelberger and Weisel say that could pass the billion-dollar mark in a decade.
Not everyone is as optimistic. Eric Mittelstadt, head of the nation's largest robot manufacturer, GMF Robotics, says he is taking a ``wait and see'' approach when it comes to service robots.
The critics' concerns are multifold. Some say that service robots will wipe out the jobs that otherwise might have been created for industrial workers displaced by factory robots.
Others question whether there really is much of a market, whether for ``nurse'' robots or for electronic butlers.
And there is also the question of know-how. Clearly, robotics technology has advanced tremendously since 1961, when Engelberger installed the world's first factory robot. Still, robots are incredibly primitive when it comes to interfacing with the outside world.
In factory situations, most welding robots still need expensive fixtures to accurately position the pieces they are going to bind together. Even the most advanced vision systems are just beginning to develop rudimentary binocular - or three-dimensional - capabilities.
To perform a task as simple as cleaning a bathtub, a robot would have to know first how to find the bathroom and then the tub, and then be able to tell whether there's a person already inside.
``The technology will be developed,'' Engelberger insists.
But how soon and at what cost will determine the extent and the speed at which the industry for service robots develops.