In an aging neighborhood of this Miami suburb, just inside the garage door of a squat cinder-block machine shop, labors a shock trooper in the Second Industrial Revolution.

Standing nine feet tall, it has a baby blue body and one bright yellow arm that's eight feet long. The trooper works 10 hours a day as an arc welder in a cramped all-woman shop. It's boss is Robbie Lozada, a divorced mother of four.

The robot welds tirelessly, for $12 an hour, doing the work of four welders who would have to be paid a total of about $60 a hour.

"I haven't given the robot a name yet, but you can bet it is going to be a female," says Lozada, 33, who six years ago was a dialysis technician in a South Florida hospital. After catching hepatitis from a patient, she switched from treating people to programming machines. Since then, she's learned how to operate three computer-controlled machine tools, along with one 2 1/2-ton computer-controlled hydraulic quasi-humanoid welder.

The Hialeah robot -- a Cincinnati Milacron T3, which stands for "The Tomorrow Tool" -- is, in many ways, just another machine. It's bolted to the floor, it's dangerously strong and it's very stupid. But, says Richard Messinger, a vice president of the Cincinnati company that makes the robot, "it just so happens to look humanish."

As such, robots symbolize -- in fact personify as no other machine can -- a high-technology revolution that has just begun to transform the way things are made in America. Robots are the missing link that enables the electronic intelligence embodied in computers to touch and manipulate the material world. Working in league with a dazzling array of computerized devices that can do everything in factories from designing new products to ordering raw materials, robots have the potential to cure what most economists and businessmen agree is the basic sickness of the American economy -- low productivity.

American workers in recent years have been getting paid more and more to produce less and less. Walter Reuther, the late president of the United Auto Workers, once said that the job of labor is to carve fat off companies. But when productivity goes flat, companies stop putting on fat. Since the early 1970s, productivity growth in the United States has stopped, and in recent years it has declined.

If productivity growth had risen in the '70s and '80s as it did in the early post-war years, economists say, the economy would be far less anemic: there would be less inflation, less unemployment, and foreign competitors would not be so successful in stealing the markets of American industry.

Robots may be able to reverse the productivity slide. Unlike American workers, robots make more and more while costing their owners less and less. For example:

In a Boeing plant in Washington state, a robot sands the wings of air-launched cruise missiles in 46 minutes, a job that takes eight hours of "diligent effort" by workers.

In a Clark Equipment plant in North Dakota, a robot welds dirt buckets for earth movers twice as fast as human welders, and last year it saved Clark $67,000.

In a General Dynamics plants in Texas, robots drill holes to help make F16 Fighting Falcons and are three to four times more productive than people.

Robots can work 24 hours a day without getting tired, bored or grumpy. They don't unionize or strike or call in sick. And robots are rapidly becoming smarter and cheaper.

"The potential of robot technology to raise industrial productivity by hundreds or even thousands of percent thus makes it a matter of the highest importance to the future of the country, and indeed to the future of modern civilization itself," says James S. Albus, a leading robot expert at the National Bureau of Standards.

Robots are predicted to cause massive change in America's blue-collar work force. In the next decade, the factory work force will be cut by 20 to 25 percent and up to 4 million jobs will evaporate, according to Thomas G. Gunn, an expert on automated factories who works for the consulting firm of Arthur D. Little. A Carnegie-Mellon University study last year found that nearly 7 percent of working Americans do the kind of jobs that robots can or soon will be able to perform.

Both the fearsome disruptiveness and wealth-generating productivity of robots have for 20 years been more predicted than realized. But the U.S. robot industry, with about 50 companies, is still an infant -- an almost insignificant part of the economy.

There are only about 4,500 industrial robots in the United States. Sales this year, hurt by the severe recession, will total about $185 million, according Eli S. Lustgarten, a robotics analyst for Paine Webber. By comparison, sales of mascara for women's eyelashes last year totaled $269 million.

"We are in the horse-and-buggy days of robots," says James Geier, chairman of the board of Cincinnati Milacron. "We will look back at these days and think of them as very primitive."

Primitive and small though it may be, the robot industry has finally begun to take off. Sales are growing at 35 to 50 percent a year, and are widely forecast to top $2 billion by 1991. At that rate, according to a recent report of the Tech Tran Corp., there will be about 250,000 robots in American factories in nine years.

The robot industry now is often compared to the infant mini-computer industry of the 1960s. Sales of mini-computers this year, surging in spite of the recession, are projected to top $8 billion. General Electric predicts that the factory automation industry, of which robots are a key part, will total $30 billion in sales by 1990. That would make it as large a part of the economy as the current truck manufacturing industry.

Robotic technology is thundering along at such a rapid pace that the Cincinnati Milacron robot in the Hialeah machine shop -- which is only four months old -- is already out of date.

"We ordered our machine and it came in 120 days later and when it arrived it was obsolete," says Sergio Morsani, owner of the Sermor Inc. machine shop, who paid $150,000 for his hydraulic robot.

Hydraulic robots have been supplanted by cheaper, more reliable electric models. In three years, robot prices are expected to drop sharply to an average of $35,000 each, according to a Society of Manufacturing Engineers study. If Morsani had waited a few months more, he could have purchased -- for less than the price of his blind robot -- a welding robot with a vision system that uses a laser to "see" the seam it is welding.

"I tell people that by the time they understand any robot, the software to run it will be obsolete," says Ken Ellis, who teaches people how to operate robots at Cincinnati Milacron.

The blurring speed of U.S. robot development is propelled by ever smarter computers. Pushing just as hard, however, is competitive pressure from the booming Japanese robotic industry. Although the United States had an eight-year head start in robotics, having built the first industrial robot in 1961, Japan is now by far the world leader.

About 14,000 programmable robots, or about 59 percent of the world supply, are at work in Japan. The Japanese government and industry have changed the economics of buying robots, with low-cost loans and tax breaks. The same machines, without the subsidies, would not be profitable in the United States.

Japan has upstaged American robot makers to the point where several large American corporations have recently thrown in the towel and begun leasing back Japanese know-how. General Electric has signed on with Hitachi; General Motors has started a joint robot making venture with Fujitsu Fanuc.

Competition in robotics, however, is not nearly as important to the health of the U.S. economy as competition -- across all manufacturing -- from nations that make things in factories using robots and other computerized elements of the so-called "flexible manufacturing system." It doesn't matter so much who makes the robots and their computerized playmates, many high-tech experts argue. But if American manufacturers don't buy them and use them in their factories, they will soon lose their competitive edge to those who do.

"Automate, emigrate or evaporate," preach executives of General Electric, which has made an enormous investment in high-tech manufacturing systems.

Says Albus, the robot expert at the National Bureau of Standards: "If we weren't in a world economy, it would be fine for Americans to leisurely introduce robots into factories at a pace that would make everyone comfortable. But the Japanese are modernizing at twice our rate. If we have slow growth and low productivity while the Japanese have fast growth, pretty soon we are going to be working for them. The Japanese will make all the products and get rich and we will not get rich at all."

Most robots in the United States have gone to work in giant manufacturing plants, with about half of them working in the auto industry -- most as spot welders. Domestic auto companies, struggling to be competitive with the robot-happy Japanese car makers, plan huge investments in robots. General Motors, for example, is committed to installing 14,000 in the next eight years.

The largest future market, however, is in small and mid-size factories. In that respect, the welding robot in Sergio Morsani's all-women machine shop -- although technologically modest -- is the robot of the future.

"If I didn't buy robots, the scope of my business would be cut 10 to 25 percent over the next couple years," says Morsani, 34, a computer buff who has spent $1 million for computerized machine tools since 1972.

"I can't find welders in South Florida. The only person who knows how to weld in my shop is me, and I don't have time and I don't want to do it. This company is going to lose the art of human welding. If we can't do it with the robot, we won't be able do it," says Morsani.

Before buying his computer-driven welder, Morsani shopped around. His search offers a glimpse of the fiercely competitive robot industry, where high-tech entrepreneurs are competing with established makers of machine tools for what will soon be a hugely lucrative market. A robot firm that comes up with the hottest technology -- robot vision, a sensitive adaptable gripper for a robot's arm, mobile robots -- could become the General Motors of the robot era. In an industry doubling in size every two to three years, a wrong move could leave today's major robot makers on the sidelines.

"The pace is breakneck, the pressures unbelievable," says Dean LaCoe, manager of training for Automatix, a Boston-area robot company. "We are working on stuff that will be in the Smithsonian."

For his Hialeah machine shop, Morsani considered a robot made by Automatix -- an upstart, three-year-old firm which just began selling the welding robot that "sees" with a laser. But Morsani gave his money to Cincinnati Milacron, a 98-year-old tool making company and the nation's second-largest robot maker. The two companies, one rooted in industrial Ohio, the other in Boston's mushrooming high-tech suburbs, represent antipodal approaches to cashing in on robots.

Cincinnati Milacron, with headquarters in an aging industrial edifice reminiscent of Charlie Chaplin's "Modern Times," has staked its future on helping manufacturers use current-technology robots. The company, which has 28 percent of the robot market, makes its robots from the ground up. They are widely considered to be reliable, heavy-duty industrial tools.

But the company does little research on vision systems, considered by experts as the key to smarter robots. Consequently, Robert Ayers, an engineering professor at Carnegie-Mellon University, says: "Cincinnati Milacron is a terribly conservative company. They make an enormous heavy-duty robot. They are doing damn little research. I think they are going to be left behind."

Automatix, on the other hand, is the darling of academics and of many Wall Street robot analysts. "It is a growing, aggressive, innovative company," says Laura Conigliaro, robot analyst for Prudential Bache.

Automatix, which has 4.2 percent of the nation's robot market, was founded by Philippe Villers, a high-tech entrepreneur. Three years ago, he attracted $6 million in venture capital from big-name investors such as Harvard, MIT and the trust that owns Arthur D. Little. Automatix makes a hybrid robot using imported Hitachi arms and Motorola computer chips. It then adds its own robot control, vision system and the software that makes the robot work.

Villers, an MIT engineer who predicts his company will become the IBM of the smart robot business, contends it makes more sense to build "value-added" hybrid machines than to construct entire robots from scratch.

"We couldn't afford to develop everything. We chose to develop a vision system and a robot control that would give us competitive advantage," says Villers. "We say our robots are smarter, not that our arms are stronger."

It's unclear, in the volatile, recession-stunted robot industry, whether the smarts of Automatix will prevail over the strong arm of Cincinnati Milacron. Both companies are painfully aware of each other. Says Villers, "Cincinnati Milacron wants to be where the iron is, we want to be where the gray matter is." Says George T. Rehfeldt, a senior vice president at Cincinnati Milacron, "Our company is not in business to hype itself. That's exactly what Villers does at Automatix."

For his part, Sergio Morsani decided to go with the older, established company because it has more service people to teach Robbie Lozada how to operate the robot and fix it when it breaks.

"But if I'd known about that Automatix vision robot, I may have bought it," Morsani says. "My next robot may be Automatix. I've got to see what happens."

In any case, Morsani has also recently purchased a $217,000 computer-aided-design/computer-aided-manufacturing system that allows his computerized machine tools to talk to each other. He plans to plug a robot into his flexible manufacturing system as soon as someone figures out how he can do it cheaply.

Explains Morsani: "By the end of this month I will have eight computers in my machine shop and I only have 12 people. I think that says it all."

Next Sunday: Robots and jobs.