2013年1月17日 星期四

Small Factories Give Baxter the Robot a Cautious Once-Over

Chris Budnick is head of Vanguard Plastics, a small injection-molding operation in Southington, Connecticut, that makes plastic fixtures, gaskets, and other “stuff no one cares about unless it breaks.” On a computer screen, placed where all his workers can see it, Budnick displays what he considers the company’s key statistic: sales divided by man-hours.
Budnick, a Yankees fan who never misses a game on the radio, calls it his company’s “batting average.” Wages are his second-biggest expense (after raw materials), and sales have been slow. Even so, the figure stands at $206.8 per man-hour, above the previous year’s mark of 201. For Vanguard to stay in business, says Budnick, the figure has to go up by 1 percent or more every year.  There’s only one way get there: produce more while working less.
That is why Budnick is now considering adding a new member to his team: a robot called Baxter. Baxter was conceived by Rodney Brooks, the Australian roboticist and artificial-intelligence expert who left MIT to build a $22,000 humanoid robot that can easily be programmed to do simple jobs that have never been automated before (see “This Robot Could Transform Manufacturing” and “Rebooting Manufacturing”).
Brooks’s company, Rethink Robotics, says the robot will spark a “renaissance” in American manufacturing by helping small companies compete against low-wage offshore labor. Baxter will do that by accelerating a trend of factory efficiency that’s eliminated more jobs in the U.S. than overseas competition has. Of the approximately 5.8 million manufacturing jobs the U.S. lost between 2000 and 2010, according to McKinsey Global Institute, two-thirds were lost because of higher productivity and only 20 percent moved to places like China, Mexico, or Thailand.
At Budnick’s shop, it’s easy to see how efficient things already are. Even though it’s a small company, with $6 million in revenue, Vanguard operates state-of-the-art automated electric presses, costing about $150,000 each, that crush plastic pellets into shapes under 1,000 atmospheres of pressure. Custom-built robots—running on tracks overhead—swing down to pluck out the finished parts and place them on a conveyor. The presses work nonstop from Monday to Saturday at noon. Sunday is for resting.
One routine job that’s still done by hand at Vanguard is packing parts. Coming off one of the presses are small, textured, plastic cups, which Vanguard sells for 2 cents apiece to a medical company to package liquid medicines. A worker from a temporary agency, earning $9 an hour, stacks the cups, then flicks a plastic bag over the stacks.
This is the job that Baxter will be trying out for when Rethink delivers a modified version of the robot, due at Vanguard for a field test this winter. Budnick says if he can eliminate one temporary worker—thus earning back his investment in a single year—he’ll buy Rethink’s robot.
“This will be a big test for them,” says Budnick, pointing to the tattooed worker stacking the cups. “Because if they can’t do that, what can they do?”
Rethink, which unveiled Baxter to wide attention in the media last September, hasn’t sold any robots yet. But it has received inquiries from scores of manufacturers, including artisanal breweries that want to see if the robot can box bottles. “Most folks are treating it as a totally new category that they should roll out in small numbers,” says Mitch Rosenberg, vice president of marketing at Rethink, which has raised more than $62 million from investors.
The ultimate goal is for robots like Baxter to take over more complex tasks, such as fitting together parts on an electronics assembly line. “A couple more ticks of Moore’s Law and you’ve got automation that works more cheaply than Chinese labor does,” Andrew McAfee, an MIT researcher, predicted last year at a conference in Tucson, Arizona, where Baxter was discussed.
Rethink brags that its robot is made of 75 percent American parts and is built in Massachusetts. The flag-waving is intentional: its target market is people like Budnick, 48, a Little League coach and former Army lieutenant who has resisted overtures to move his factory to Mexico and refuses to buy machine molds from mainland China because “they are not free people. They don’t vote.”
But Budnick won’t buy Baxter out of patriotism. When Rethink executives first brought the robot to Vanguard for a demonstration early in 2012, he and his father thought Baxter looked like a relative weakling compared with their imported Austrian presses, some of which have been operating for 100,000 hours.
“They came in and said, ‘Well Mr. B, what do you think of that?’ And my dad said, ‘Well, you know, you guys have to improve it 100 percent,’ ” says Budnick. “And all the engineers’ jaws dropped.”
Baxter comes with two arms, a vision system, and 360° sonar (which it uses to detect people nearby), but for the cup-stacking job it will also need a specially designed gripper, which Rethink is now developing. Rethink is also developing software so that the robot can communicate with other machines, such as a conveyor belt, telling it to move forward or stop.
So how important will Baxter really be to Vanguard? Budnick couches his answer in baseball terminology. “Baxter is a potential double,” he says. “Maybe a home run if it can use both its arms.” 

30 comments. Share your thoughts »

If Baxter is purchased by an employee, said employee is paid 40 hours a week but Baxter works  120 hours a week, both employee and employer have just entered a win-win situation. Employee is responsible for maintenance of Baxter not his programming. Employer provides programming for Baxter's task.  Thus you have employment for people and you get to advance  U.S. based manufacturing . Employers and employees are limited to the amount  of Baxter's they may own  and operate. This will all come in very handy when we start mining  and assembly and manufacturing jobs in space. Have a nice day.

The Baxters of the world will become ill. Circuit boards and components fail over time and the mechanical parts will wear out. So the factories will need to carry spare parts or spare robots. So the $22K cost may not cover the whole expense of a one-for-one replacement.  If the electrical system fails then the robot is just a pile of do-do. The temp could keep working if other steps in the process were also manual. Plus I wouldn't be surprised to see some sabotage either.