Rethink Robotics to increase presence in Australia, up R&D spending on Baxter

This month Rethink Robotics received a substantial chunk of new funds from investors, led by GE Ventures. Brent Balinski spoke to the company’s CMO Jim Lawton about plans for investing in expanding into newer markets and improving its Baxter industrial robots.

Robots, for industrial and other settings, have never been more popular.

 

You can turn back to mid-last year and the International Federation of Robotics’ last annual audit of global industrial robot sales and see that numbers have never been higher, up 12 per cent on the previous result.

 

Last year saw consumer robots released for everything from cleaning cat litter (Litter-Robot) to connected, intelligent gardening. And Dyson even chose 2014 to release its long-awaited robot vacuum.

 

Things don’t look to be slowing down this year, with a bumper beginning to 2015 for robotics startup funding. On January 22, according to GigaOm, funding worth $US 51.9 million had been raised so far. Small beer for other industries, sure, but pretty healthy considering a total of $US 341.3 million raised for the whole of 2014.

 

Among those receiving a cash injection was Rethink Robotics, makers of the Baxter industrial robot, securing $US 26.6 million in Series D financing.

 

Plans for the money include investing in the upgrades to Baxter’s software – the key to improvements in how quickly and accurately Baxter works – and increasing Rethink’s presence in markets such as Europe, Asia and Australia.

 

“The company officially started selling Baxters in September 2012, we started shipping in Q1 of 2013, and really the last two years have been focussed on ‘let’s make sure everything is the way that we want it to be’,” the company’s chief marketing officer, Jim Lawton, told Manufacturers’ Monthly.

 

“And when we got to that point, really over the summertime and the fall, we said ‘okay, now’s the time to scale.’”

 

Rethink’s robots are in hundreds of factories in the US, with the company choosing their major customers close to Rethink’s Boston headquarters. There are much fewer Baxters, distributed by Training Systems Australia, in this country.

 

Factory robots have been around since the 1970s, with their introduction led by General Motors.

 

Their use is still dominated by the auto sector, which, according to Rethink’s founder Rodney Brooks (an Australian native) takes in at least 70 per cent of the total.

 

As anybody who has seen heavy-duty robots, such as those used in car assembly plants, will tell you, these robots are nothing you want to stand in the way of. They are usually kept in cages to prevent this from happening.

 

These fixed, dangerous, expensive machines also tend to require large amounts of time spent on coding, with further changes necessary if things need to be moved around.

 

The Baxter was designed (in top secret conditions, over almost five years) to be everything old-school robot help wasn’t: collaborative, able to be quickly programmed by a layman, comparatively lightweight, and safe to be around workers.

 

“We built a robot with a compliant arm – I mean it’s literally made out of the springs, the seven joints have a spring in each one of their joints,” explained Lawtown.

 

 “So it’s got a certain amount of give, so when my hand bumps up against the wall, the wall, essentially is not moving. But my hand will kind of push back a little bit, and that prevents me from being hurt and it prevents me from damaging the wall.

 

 

One of Rethink’s goals is to develop the kind of robots that can be used in all of the tasks that are not yet good candidates for automation today.

 

And the company says something like 95 per cent of manufacturing tasks can’t be automated the old way, and this is driving the demand for flexible robots such as those by Rethink and other companies (including Universal Robots, which last year expanded into Australia.)

 

Baxter’s arms are able to move with seven degrees of freedom, like human arms, and the “face” on the robot’s screen gives co-workers cues as to how to robot will act. It will look at an object before it is about to pick it up, for example, and will display a confused expression if there is a problem while trying to carry out a task.

 

Important for the Baxter and other collaborative robots will be the ability to act more and more like the humans they are working alongside. Lawton sees the improvements as coming in the areas of how well the robots engage, learn, adapt, resile and sense in their environments.

 

Unlike the heavy, fixed factory helpers of the past, Baxter does not require some 200 hours of coding to program an operation, says the company.

 

And though its mixture of position and force-controlled robotics are sophisticated, great efforts have been made to keep the trickiness behind the scenes and make programming the Baxter as intuitive, easy and quick as possible.

 

“So we’ve kept that complexity built into the product and all they have to worry about is moving the arm in a way which you want it to go and have the robot remember that and it’ll just do that over and over again,” said Lawton.

 

Baxter was the creation of celebrated roboticist Rodney Brooks and colleagues from MIT. Originally launched as Heartland Robotics in 2008 the company became Rethink in 2012.

 

Brooks had previously founded iRobot, the maker of the popular Roomba vacuum, in 1990.

 

A manufacturing company that sells to manufacturers, Brooks’ vision of a “new category of robotics” came about through the experience of outsourcing to China. While looking for a manufacturing partner to iRobot in 1997, the situation of outsourcing to chase cheap labour seemed an unsustainable one in the long term..

 

“Even in the late 1990s, if you were doing sewing for toys [for example], that had already moved to Vietnam,” he told an NPR interviewer last year.

 

Also, what could be automated was limited and doing so was cumbersome.

 

Heavy, fixed robots, aside from being dangerous and expensive, are difficult and time-consuming to adjust if this is needed.   

 

“What you’re doing, also, by putting those cages in place is number one you’re making it impossible for a person to interact with a robot to keep them safe; but to the extent that you have ideas about ‘what if I could change this and make it a little better?’ – all the continuous improvement, the six sigma initiatives that a lot of manufacturers have, are essentially frozen in time,” said Lawton.

 

“They’re halted because the only way to make those changes would be to hire someone to come in, reprogram the robot; if it means that the cage has to move in a different way, they have to unbolt it and re-put it back down.”

 

The programming of Baxter is at the heart of its usefulness. A comparatively cheap robot, its price kept down by inexpensive parts, a unit is available here for around $35,000.

Rethink aims to make performance better and better by regular upgrades to its Intera software platform.

 

The Intera 2.1 upgrade in last April, for example, doubled the speed and increased precision of Baxter.

 

Another upgrade rolled out in November was the addition of a Robot Positioning System feature (part of the Intera 3.1 upgrade) allowing the machines to work considering the relative positions of different coordinates in an environment, and adapting more quickly to different tasks. If a Landmark (an environmental marker) moves, the robot will adjust accordingly.

 

“All you do is you show it what it needs to do and once you show it it does it over and over again,” said Lawton.

 

“And the reason that we did that is to make it impervious to the change in variability that happens in the environment, but also happens in a way that’s seamless to the end customer.”

 

Though the Baxter’s sale pitch includes messages about making workers more productive and freeing them from boring tasks they’d rather not do, there are frequent questions about whether or not collaborative robots will take more and more manufacturing and other jobs.

 

An influential study from Oxford Martin School in 2013 suggested nearly half of US jobs could be replaced by computerisation over the next couple of decades.

 

In Australia, there are similar fears about the effect of sophisticated automation on jobs. In December the Department of Industry released its Australian Industry Report 2014, finding up to 500,000 white collar jobs could be replaced through automation.  

 

And recent super-scale automation projects, such as those at ports and mines, are evidence to some that as machines get better, economies will employ fewer workers.

 

Rethink says that fears are, to some extent at least, based on whether people are “glass half-full or glass half-empty” types.

 

“I would say that no doubt there are jobs today that are not going to be around in the future,” offered Lawton. “Just like 20 years or 100 years ago there were jobs around that don’t exist today.

 

“But I think anybody who says ‘there isn’t going to be jobs impacted’ is not being entirely truthful.”

 

Lawton believes the future is an exciting place for manufacturing, and will be transformed by technology. Managing that transformation is important, and enabling the skills and training to adapt will be crucial for policymakers and firms.

 

“The other part of it is we’re starting to see a lot of manufacturers say ‘what happens when I take some of these interesting technologies – like the industrial internet, the internet of things, additive manufacturing collaborative robots – and you start to come together in very creative ways,” he said.

 

“I think we’re going to see this kind of convergence of data and machines and software that will allow us to do some things that we can’t even think up today.

 

"So I think we’re going to go through this transformation of some jobs are going to go away, a lot of jobs are going to get created – I think the real challenge from a societal perspective is how do we help these people that could potentially get caught in the transition.”