Tending to the collaborative robot push: an interview with Rethink Robotics

Rethink’s latest offering adds to the crop of new, easily-trained robots that some believe will revolutionise the way factories operate. Brent Balinski spoke to Rethink’s Jim Lawton about the robot Sawyer and the current excitement within the field of collaborative industrial robots.

Going where other robots won't

The worldwide market for industrial robots is booming, with electronics ranking high among drivers of growth by sector and China leading the way among geographic drivers.

It makes sense that Rethink’s newest robot, the Sawyer (announced last month) is aimed at both electronic manufacturing and the Asian market.

“We're aiming at the electronics assembly industry — and most of that will be in Asia,” company founder Rodney Brooks (pictured below) said this month.

Sawyer, a one-armed collaborative robot, runs on the same Intera software platform as Rethink’s first robot, the Baxter. It also features the same “face” on its screen (also used in programming the robot, and for telling nearby workers what the robot is about to do).

However, there are differences in what the newest addition to Rethink’s robot family can do. According to the company, these will help it attack some of the remaining nine tenths per cent of manufacturing tasks that have been hitherto un-automatable.

Sawyer suits certain tasks that Baxter is not ideal for.

“It is higher precision, faster, it still has seven degrees of freedom, but its precision allows it to get into and perform tasks we couldn’t do before: certain categories of machine tending tasks, FPGA [field-programmable gate array] programmers, ultrasonic welders, places where you need to be able to relatively finely go in and be able to insert a component leveraging the mechanical compliance of the robot arm,” said Jim Lawton, Chief Marketing Officer at Rethink, told Manufacturers’ Monthly.

“The Sawyer is a more precise robot and a faster robot and so for tasks that involve cycle times that are of a higher speed, the robot is able to perform those tasks.”

Baxter has been put to uses such as line-loading and packaging, though the improvements in dexterity are aimed at covering fiddlier, though still repetitive, tasks.

The smart money says…

And where Baxter has only been sold outside of the US to corporate research centres and research institutions, its younger “brother” in the family of robots is designed to be a global product.

Sawyer was launched on March 19, with limited availability scheduled for the mid-year and general availability later in 2015. Its development began in late-2013.

There appears to be a lot of faith in what the robot and its predecessor will eventually offer the world’s factories. At the end of last week, Rethink announced its final chunk of Series D venture capital investment: $US 13.4 million from Wellington Management Company, with the entire round worth $US 40 million for the Boston-based company.

This makes a total of $113.5 million in financing raised, despite Rethink having only sold “hundreds” of Baxters so far, reported IEEE Spectrum the day of Sawyer’s launch.

The company’s first robot was launched in 2012, and Rethink plans to create a family of robots based on the same Intera program as Baxter, each for different types of tasks.

For its first generally available robot, there is a huge potential in servicing manufacturing’s “inconvenient truth”, Lawton has written.

Machine tending is a drain on factory productivity, says Rethink. Repetitive tasks such as worker operating a press brake or PCB tester – holding a part in place for a period of typically 30 to 90 seconds – not yet candidates for automation, despite being often very boring and simple.

Lawton compares much of this sort of work to watching “paint dry”, with workers probably preferring to contribute something more creative and personally fulfilling to their workplace.

Sawyer is able to successfully slot into such situations, said Lawton.

“What you don’t want to do is put a really expensive robot in front of a really expensive piece of equipment, so it has to be relatively affordable [Sawyer has a base price of $US29,000],” said Lawton.

“But what you also don’t want to do is put an inexpensive robot in front of a really expensive piece of equipment and because it’s slow, it slows down the pace of the really expensive piece of equipment.

“So, the robot has to be at least as fast so that it can keep up with what the piece of equipment is. And if they’re inexpensive you can start to double them up if you’ve got capacity constraints and have floorspace to do that. You just can’t slow down the line because of the speed of the robot. Sawyer’s faster than Baxter as well.”

Lighting the way in factories

Another thing that separates the two robots is the higher-resolution Cognex camera, positioned in the robot’s “wrist” rather than what you might call its “palm”, where it is less likely to be occluded during tasks. It also includes a built-in light.

“Ambient lighting is a real trick for vision in a lot of manufacturing environments,” said Lawton.

“You can get the task to work in the daytime and then when the sun goes down it doesn’t work anymore because the ambient lighting has shifted.”

The original version of robot was introduced to “validate” a hypothesis behind a new kind of robot, according to the company. As noted many times before, conventional factory automation has been unsafe to be around, heavy, hard to program, inflexible, and generally expensive.

The company’s first collaborative robot was designed to see if a new way of doing things could handle what Boston Consulting Group estimated were the nine-tenths of manufacturing tasks that couldn’t be automated in traditional ways.

“The robots are smarter, they apply more common sense than what traditional automation does, and they’re able to do the tasks that automation hasn’t been either cost-effective or practical to go after,” said Lawton.

“Now we just need robots that can do other chunks and so that was one of the reasons we designed Sawyer.”

According to the company, most of this year’s allocated volume for sales has already been spoken for.

Australian enquiries regarding Sawyer have been strong, said the CMO, crediting this to Baxter’s high profile and “the experience that companies have had in the research and academic environments [regarding] what collaborative robots can do.”

Shipping out here is scheduled to begin early next year.

The co-bot business is booming

The collaborative class of robot is anecdotally experiencing rapid growth. Hard figures are not yet available for global sales, the International Federation of Robotics told Manufacturers’ Monthly, due to the newness of the category.

The same week as Rethink released Sawyer, its competitor Universal Robots of Denmark released its UR3 machine. It has a similar payload (3 kg, Sawyer’s is 4 kg), the ability to be programmed by a non-expert, and a pitch towards those wanting the automation of relatively dexterous tasks (“gluing, screwing, operating tools”).  

And just last week, Switzerland’s ABB joined the burgeoning “co-bot” market with its YuMi, unveiled at Hannover Messe.

Lawton said there was definitely a lot of excitement in the field of collaborative robots. A couple of years ago there were two companies offering such robots, he said, but now there was roughly a dozen.

He predicted that those running factories would see more and more options for person-friendly robot help. Elsewhere, Lawton has predicted that in perhaps to five to 10 years, no successful manufacturer will be without a collaborative robot.

“I think more generally speaking there are going to be a number of successful players in the collaborative robot space,” he said.

“We of course think Rethink Robotics is going to be one of them… The number of robots going to be sold that are collaborative robots is rapidly going to dwarf the [number of] legacy robots.”

 

Images: supplied