Here is the Kurt Konolige interview. Dr. Konolige is a Stanford professor and a senior researcher at Willow Garage. Willow Garage is a startup robotics company that is focused on aggresively promoting and actively encouraging the long-term development of robotics:
Willow Garage is developing robotics hardware and software and has developed several unique robots, including the PR2
Question: Willow Garage is a unique organization. Wouldn’t it be more cost-effective in the short run to develop proprietary intellectual property instead of emphasizing open standards?
Answer: We actually are developing commercial products, although that isn’t currently our immediate goal. There are two aims to our business model. First and foremost, we want to accelerate the development of robotics technologies by creating a series of open standards and seed the field. But we also want to create useful and profitable product lines, and these product lines may incorporate proprietary IP.
Question: Are there any robotics devices that are close to commercialization?
Answer: Yes. We have developed a telepresence platform called the Texai. This partially autonomous robot can be remotely operated and controlled, and we are building 25 of these robots. Although we are not currently selling the Texai robot, we are amenable to doing so given sufficient demand.
Question: Tell us about the PR2 robot development platform.
Answer: We are giving away 10 PR2 robots to organizations that can use the robots for further development. We have received an overwhelming response – 120 letters of intent to respond, so the robot is quite popular. The PR2’s arms are actually quite sophisticated, so they can be used around people. We also have a full set of software to perform tasks such as moving around, perceiving objects, and manipulating objects. The important thing about the PR2 is that the hardware is open, and the software is 100% open source. The PR2 will make code sharing in the community a lot easier, allowing all of us to stop reinventing the wheel.
Question: So will you ever sell the PR2?
Answer: The initial plan was not to sell them, because this is primarily a research robot. While the PR2 has generated considerable interest, we are not announcing any plans to sell them at this time. The primary purpose of the PR2 is not to generate revenue but rather to expedite the development of technologies necessary for a robotics industry.
Question: Will there be a PR3 robot?
Answer: At this point, we have just gotten the PR2 robot out, so we are focused on that. As the PR2 Beta Program robots are used in the field, we will learn a lot about the system. I would like to see a PR3 robot developed.
Question: To what extent is the robotics industry constrained by a dearth of standards?
Answer:The biggest problem that the robotics industry faces is the gap between what is needed and what can be delivered. At this point, the robotics industry is hampered by a lack of good, safe, flexible arms. If robots were competent enough to recognize and manipulate objects in real time that would constitute a major advance. Developing robust arms will require advances in both hardware and software.
Question: Why do arms present such a formidable challenge to roboticists?
Answer: Industrial arms are positionally controlled and quite powerful, but they aren’t safe enough to operate around humans. Arms need to be compliant, and capable of a wide range of motions and movements. Fortunately a number of excellent arms are being developed.
Question: Is robotic vision improving?
Answer: Vision is another daunting challenge. Positional control is adequate for a tightly controlled assembly line, but robots need to be able to discern objects in their environment, and make continual adjustments based on feedback loops. At this point we lack the algorithms and the processing power needed to do this in real time.
Question: How is Willow Robotics actively promoting these open standards?
Answer: We are developing a robot operating system called ROS (Robot Operating System) and ROS is available on both Linux and Mac operating systems. ROS carries a BSD license, so it can be used for any purpose, including commercial, without restrictions. There is already a thriving community of ROS users and developers around the world. One of our goals is to grow a ROS community similar to the Linux OS community.
Question: To what extent can the PR2 be improved? What new features could be added?
Answer: The PR2 is a very competent and versatile machine, and we are pleased with the design. But the pincer hands that we use, although surprisingly useful, do have limitations. So we could potentially add hands to future models. The next big step for the PR2 robots is to get them out into the robotics community. We expect future design improvements to be guided by the research and application directions taken by the community.
Question: How many researchers are working at Willow Garage? Is Willow Garage a corporation or a nonprofit?
Answer: The company is a mixture of researchers, software development people, hardware development people, and manufacturing people. There are currently a little over 50 full time employees. The largest component is the software development team, which comprises around 20 people. We are a corporation but are dedicated to open source, and we are not constrained by the need to generate short-term profits.
Question: What is the first potential market for general-purpose robots?
Answer: There are a number of dangerous tasks for which robots would be preferable to humans. However, getting sufficiently competent robots is the challenge. But there are places such as packing centers, bakeries, and nursing homes where even simple robots would be useful. We could see robots deployed in those areas in the next decade. In the long run, robots will be used in virtually every industry and will become ubiquitous.
Question: Who is funding Willow Garage? When will it become profitable?
Answer: Willow Garage is a privately funded company, owned by investors with a long-term perspective, who value creating positive impact on the world as much as generating revenue. We are only three years old, so it is a little premature at this point to think about generating profits.
Question: Is Willow Garage doing direct AI research?
Answer: We are working on perception for object recognition and mapping. We are also working on task level planning and manipulation planning and human/robot interaction research. But we are not directly researching artificial general intelligence.
Question: If no breakthroughs in artificial intelligence occur, how capable can robots become?
Answer: That would depend on how one defines “artificial intelligence” and “breakthroughs”. I see progress in AI as being incremental, rather than the sudden creation of a sentient machine. But AI is making steady progress in all manner of areas, and this is leading to robots that are increasingly capable.
Question: How much progress can reasonably be expected in the robotics field during the next decade?
Answer: The robotics field is clearly progressing much faster than it was ten or twenty years ago. Sensors are getting better, computing power is exponentially increasing, and the hardware is improving. The aim of Willow Garage is to further expedite improvements in robotics, with the goal of creating a thriving robotics industry as quickly as possible. I don’t know if there will be a robotics “killer app”, but as robots become increasingly capable, the pace of progress will further increase. In 2020 we will look back on current robots and see them as quite primitive.