background image

Brooks’ Baxter Bot

With the recent announcement of ReThink Robotics’ Baxter (see link below if you haven’t seen it), I received a lot of questions about how their robot may impact what Hoaloha Robotics is doing.

http://vimeo.com/49683995

First, let me say that this unveiling reflects the tremendous progress of robotics technology. Dr. Brooks is one of the preeminent pioneers in this field, having formerly served not only as the head of MIT AI lab, but also as one of the founders of iRobot, creators of the popular Roomba vacuum cleaner bots and the military Packbot. Brooks left iRobot and MIT around the same time I left Microsoft.

While the precise details of what he would build weren’t known, Rod was open about his intent to move industrial robots out of the factory where, for safety sake, they are fenced to avoid dangers to humans; and instead able to operate alongside people. This required not only “re-thinking” the safety issues (since industrial robots typically have little to no ability to sense or adjust to human presence), but also simplifying programming of the robot as well as reducing the costs of such a robot.

To be fair, a lightweight industrial arm that can be trained by example and is “compliant” or safe around people isn’t new. I observed such capabilities demonstrated by Kuka Roboter in 2006, using their Lightweight Robot, designed by German aerospace company DLR. Similarly, one of Kuka’s main competitors, ABB, also demonstrated a dual-arm light industrial robot, dubbed FRIDA, which is in pre-production trials. Likewise, other companies like Motoman and Universal Robots have also recently shown robots in this category.

However, there are several things about Brooks’ approach that are unique. First is the price, estimated for Baxter at $22,000. This is roughly a third of most competitors are in this category. Further, Brooks promises to make the platform open for software development. This has been traditionally unheard of in the industrial robotics sector because of the potential dangers of the robots improperly programmed, but also in enabling vendors to maintain better control over customers. In contrast, Brooks’ move now opens the opportunity for a new platform for universities focused on advancing dexterous manipulation software techniques, in addition to the more expensive options like the Willow Garage PR2 (though WG gave many away to key universities). Through this we are likely to see advancements that move beyond the light industrial sector into the home domestic sector, long the dream portrayed in film and fiction.

The leading question is still how does Brooks’ announcement impact what Hoaloha Robotics is doing presently? Baxter’s price, lack of a mobile base, weight, and programming requirements still place it beyond the near-term target for assisting seniors in the home. Our price target is more between $5,000 and $10,000 and autonomous mobility is essential in our value proposition and user experience. So I believe we are still at least 5-10 years, likely longer, before robot arm technology evolves to a point that we can add arms to our robots, though Brooks’ work confirm that it is a reasonable expectation for the future.

Brooks’ robot also sports a “face”, something that has been a key design point for Hoaloha’s robot from the beginning. While research confirms that people tend to regard robots as “social actors”, and Brooks’ prior association at MIT (where former students like Dr. Cynthia Breazeal is a leader in the field of social robot design) gives good reason to for him to have considered this, it was nevertheless a pleasant surprise. In his description of why he gave Baxter a face, Brooks talks about how the face can be used to convey important attributes that impact human-robot interaction. This is still an unsettled debate even among avid roboticists as to whether robots should rather be designed to be more like appliances. Obviously I side with Brooks. Baxter’s face helps illustrate the benefits of how social attributes are essential for human-robot interaction design, somewhat like the early Xerox Star’s graphical user interface did for Windows and Macintosh.

Further, Brooks’ is clearly not targeting (at least initially) for the market Hoaloha is focused on, i.e. empowering the senior population. Instead he is working to get robots into light industrial or small business venues. Brooks hopes to have it help keep manufacturing in the U.S. rather than go to cheap labor in other countries. One can also easily see why Bezos was an early investor. Paired with Amazon’s recent acquisition of Kiva Systems’ warehouse robots, Baxter could enable virtually full automation of Amazon’s order processing, making them an even more potent merchandiser of consumer goods.

In being clearly focused on a market where robotics can be applied, Brooks does what few others in community have done so far ; that is nailed a strong value-proposition (something he also did while at iRobot). To do so requires not only addressing a clear market need, delivered through a great user experience, and at the right price.

In the end, Baxter does not (yet) directly impact Hoaloha’s short-term designs or objectives, but it does demonstrate the potential of where we may be able to take our technology in the future. Brooks’ bot does one other important thing as well. It increases the awareness of how robotic technology is increasingly accessible and affordable, creating a positive example and expectation. I somewhat think of the parallel of how mini-computers paved the way for personal computers, substantially lowered the costs and complexity, opening up computing to smaller businesses, and ultimately opening doors for the personal computer market.