Willow Garage had it's first open house on Oct. 16. It was geared to the researchers that stopped by in Silicon Valley on their way back home to Japan and other parts of Asia after attending IROS, and I had the privilege to join them. There were about 25 participants.
(Photo: WG CEO Steve Cousins shows off one of the PR2 prototypes to the researchers)
What was new to me is that they now have technicians working there building the PR2s. They are stocking up enough parts to eventually build 25 PR2s, and according to the production manager, the short-term goal is to complete 5 of them by the end of this year. Willow Garage will "issue a Call for Proposals -- allowing researchers worldwide, from academic, non-profit and for-profit organizations to apply to secure a PR2 development platform." Ten PR2s will be available for free under this plan. The company will use 10 of them in-house and 5 will be for stock.
It takes about 1,300 types of parts and 10,000 parts overall to build one PR2 and they expect it to take anywhere between 100 and 200 man hours to complete one. The learning curve is pretty steep but once "things start rolling" they should be able to build 2 robots per week with 8-10 technicians, said the production manager.
They are doing extensive and rigorous testing to make sure things don't break easily. High quality will mean less maintenance after they ship the robots. PR2 is meant to last 3,000 working hours. It is designed to be "robust so that buggy code doesn't destroy the robot," said Keenan Wyrobek, Co-Director of the Personal Robotics Program. And even if something does go wrong, the robot is modular. So for example, you could replace the whole arm in 20 minutes, according to Keenan.
Also during the open house, software lead Brian Gerkey made a presentation titled "Towards a Robot App Store." The analogy is the cell phone market. If you build the right infrastructure, there will be an explosion of applications. To accomplish that in the robotics world, he and his company believe that the "core components of robots should be OPEN." The core system is not perfect so everyone needs to contribute and researchers need to see and change how things work.
Willow Garage's message to the research community - Let's share experimental code so that others can replicate, refute or extend results.
Economic value is at the end = applications. Then there was a question from the audience, "Where do you close the code?" Brian's answer was "I don't know where the line is, but we haven't hit that point yet."
