The company is building open-source robots and giving them away to university research groups in an attempt to fuel quick improvements to the operating system and rapid build-out of applications to run on it.
The company plans to deliver 10 robots to US universities by the end of the year, says Willow Garage president and CEO, Steve Cousins. “We might deliver more robots later, maybe up to 50”, he says. The company will release its open source software at the same time.
“You have to have the device to start inventing applications,” Cousins says. “Usually the first step in building a robot is to create the hardware and then the software. Now we give a ready made platform for research groups to start from.”
He compares the development of personal robots to the revolution of the personal computers: When the first ones came out, nobody knew what to do with them. But then the development got started.
Privately-funded research lab Willow Garage was silently founded last year by Scott Hassan, the chairman of the board, who helped Larry Page and Sergey Brin to develop Google’s technology back in the ’90s. Hassan was also founder of eGroups, a group email messaging company bought by Yahoo Groups in 2000 for about 450 million dollars. And presumably those funds are at least partly responsible for powering Willow Garage, although Hassan won’t say where his funding comes from.
Besides personal robots to be used in homes and offices, Willow Garage is developing unmanned cars and autonomous vessels. The company aims to expand from its current 20 employees to 60, as soon as possible, says Cousins.
According to Cousins (pictured left), it was Stanford University in Palo Alto that came up with the idea of making robots available to research groups. The university has been developing personal assistant robots in its Stanford Artificial Intelligence Robot (STAIR) project and has already managed to build a robot that moves on wheels and is capable of performing simple tasks.
STAIR researchers wanted to develop the robot further but funds were lacking, Cousins says. Willow Garage then gave Stanford almost a million dollars last fall to develop open source software for the personal assistant robot.
“Lack of integration is a problem: almost every research group creates it’s own hardware and software,” says assistant professor Andrew Ng, who leads the STAIR project. By having the research community move toward a standardized operating system, ideas and inventions can be shared, he says.
“Now one group works on a robot that understands speech, the other one on robot learning from its experience, and a third one on a robot that can navigate around the building. With the STAIR project we gather these separate fields and integrate them into a single robot. Revolution in robotics will come when we can build a single robot that can do a lot of different things,” he says.
Cousins says that the robots will probably be leased to the universities with conditions such as if the robots are not used in research, they have to be returned. And what does Cousins hope the universities will deliver? “I hope people will publish a paper about the inventions, make the code available to open source, and somebody else can build on top of their results,” he says.
The robot Willow Garage provides to universities will be PR2, the second generation of Stanford’s personal assistant robot PR1. Willow Garage will design the hardware and organize the manufacture, and is also involved in creating the software. PR2 will move around on wheels and performs tasks like picking up objects and opening a soda can on its own. However, it has to be controlled by human hands.
“The challenge is to build a robot that does the chores on its own, without teleoperation,” Cousins says.
The whole venture seems to be a very long-range investment. You don’t hear Cousins talking about market share or ROI. He says he’s not worried about somebody else profiting from PR2 and the open source software provided. “If you look historically at technologies that have had a huge impact on the world, like personal computing, there’s an exponential growth curve for a while. If robotics is one of those things, there will be an ecosystem of companies created around personal robots. Being in the center of that explosion, you don’t have to own everything to be successful, you don’t have to patent everything you do.” Although Cousins does say the company may patent parts of the hardware at some point.
Cousins says that the popularity of Roomba, the vacuum cleaning robot made by iRobot, which has sold millions, proves that there’s a market for robots. One possible use he points to for the personal robot is to help the elderly and disabled.
So when will we be able to buy a robot in a shop for a reasonable price? “Maybe never,” Cousins says. But that doesn’t mean they can’t be pervasive. Cousins pictures robot rental services, for example. “Maybe you can lease a personal robot to clean up your house when you’re at work. If the robot sees something it doesn’t know what to do with, it calls somebody in a call center, which might be placed anywhere in the world, sees what the robot is seeing, says don’t touch that or let me control the hand and do it manually.”
Before joining Willow Garage, Cousins was a senior manager of the User-Focused Systems Research Group at the IBM Almaden Research Center in San Jose. He also worked at the Advanced Systems Development Laboratory at the Xerox Palo Alto Research Center.