A glimpse of the future: Baxter the robot
Rodney Brooks, a thought leader in robotics and entrepreneurship, came to Pittsburgh the other day. Carnegie Mellon, courtesy of Matt Mason and the Robotics Institute, hosted Rod for the public unveiling of his new talk, “A new class of industrial robot.” Rod is the founder, Chairman and CTO of Rethink Robotics. I featured Rod and his new company in a NewVenturist post back in June, 2012. Rod is better known, perhaps, for his role as co-founder and past CTO of iRobot Corp (Nasdaq: IRBT), where he spearheaded development of many of the company’s military robotic products and, eventually, the vacuum cleaning Roomba®.
The talk was to introduce Baxter, the Rethink industrial robot that is “unlike any others,” Rod assures the crowd of a few hundred students and faculty. Rod starts by explaining two themes that are relevant to his talk and to the state of robotics in today’s economy:
- The need for human-centric robots; and
- Will we always manufacture consumer goods by hand?
Rod outlines how he came to understand the relevance of the two points. Like computers were in the past, behind glass walls, robots were isolated from people because they were dangerous to people and people interfered with a robot’s activity. But, if you look at where robotics is today, and I have written about this in my robotics series, we have Intuitive Surgical, Kiva Systems (which was acquired by Amazon), and iRobot – all of which have products in the market. And there are many earlier stage companies that either have or soon will have products in the market. Since 2002, the military has gone from 0 to >12K ground robots, and during the same time robots in people’s homes has gone from 0 to >8M. What has caused the plethora of robots?
Rod points out that several factors: the price of computation has gone down; the price of sensors has also gone way down; computer vision capabilities have increased exponentially while the price of the technology has gone down significantly; and robots can now do tasks for people that are useful.
Rod learned a lot from his years at iRobot. One of the key lessons is that price matters. The Roomba was not the first robotic vacuum to hit the market. But the Roomba was the cheapest, by far, at less than 10 times than the price of the competition. The second lesson was that “Users are not you.” By that he means that users ≠ engineers. “Real robots have real users,” he explains. Using the Roomba as an example, Rod explains that the user interface of the first release of the vacuum was difficult for users. The company received a lot of complaints. Over time, the company reduced the user interface to one button that says, “Clean.”
Another example is a particular military robot deployed in Irag and Afghanistan. The user interface was terribly complex and it took hours for users to gain control of the robot. By introducing a game controller, similar to the ones the young soldiers used for their Xbox systems and other games at home, the time to master the robot was reduced to a few minutes.
Rod moved on to his question about how long will we continue to make consumer products by hand. His argument that Chinese factories and other offshore assembly factories are not sustainable in the long-term is reinforced by the data that the hourly wage of Chinese factory workers has gone from $.90 per hour to $3 per hour in the last few years. Rod believes that while we may never see robots build complex products like iPads, robots can be used to help manufacture many products. That way, Rod, although born in Australia, sees that we can bring manufacturing back to the US. He cites reasons why US manufacturing makes business sense, including that IP will be easier to protect, it will take less time to get from the factory to the store, and transportation costs will be much lower.
So, if we follow Rod’s premise that “We need robots to make stuff,” then we need “A new kind of robot.” And that’s exactly what Rethink has developed in Baxter. “It may be the Commodore 64 of industrial robots,” Rod admits, “But someone has to do it first!” Rod believes that to create the new robot that is needed, he had to literally re-think robots.
The Rethink logo depicts a human character next to a humanoid robot. To Rod, that says it all. He believes that the next generation of industrial robots, of which Baxter is the first, must be able to interact with humans in a simpler, natural way. And the robot has to be easy to use, with “no integration, no programming, and very capable with a wide range of simple tasks.” The robot also has to “know about people, be intelligent, and be safe.”
Rod is proud of the “Made in the USA” label on Baxter. The robot was designed to be manufactured in the US via a distributed supply chain. Rod and his team went to particular areas that specialize, such as Penn State for the new metal used in the robot’s hardware.
Rod outlines what used to be important for companies seeking to use robots for solutions: “Precision, repeatability and speed.” Then he contrasts those qualities with what Baxter brings to the table, or manufacturing floor: “Adaptive, flexible, ease-of-use, safety, inexpensive, and high throughput.” Baxter, Rod insists, “Knows what you mean and does what you want!”
Baxter is cute. He/she has these super long arms with more joints than we do so he/she can do complex things like picking up and putting down. He/she has a square face, a screen, with cartoon like eyes and eyebrows that can express emotions. There is an endearing quality to Baxter that the Rethink team built into the robot to make human-robot interaction simpler. Training is much simpler than with other robots, Rod explains to the audience, showing us with videos how it is done. Baxter “watches” and learns real-time as his/her “trainer” shows what to do. And when the electric plug is pulled, Baxter kind of goes to sleep.
Rod will be back in Pittsburgh on October 22 for the Robobusiness Leadership Summit where he will unveil the actual Baxter robot, live and in person.