October 25, 2010

Davide Faconti of PAL Robotics is interviewed by Sander Olson

Here is the Davide Faconti interview by Sander Olson. Dr. Faconti is the project leader for PAL Robotics, a Spanish company that is funded by investors in Abu Dhabi. PAL recently introduced the REEM H2 robot for the commercial market, and has demonstrated the REEM-B robot. The company is currently working on the REEM-C humanoid robot, which should become available next year. There are a now a number of humanoid robots in existence or being designed, including REEM-C, Virginia Tech's Charlie Heavy, Honda's Asimo, Korea's Hubo 2, and Iran's Surena robot.

Here is a link to their life in the robotics lab blog

Question: The PAL project is focused on humanoid robotics. What is the current status of this project?

Answer: The PAL project began about six years ago with 4 workers. Today there are about 25 researchers, and we are focused on rapidly improving these humanoid robots, in particular the REEM robot. We are headquartered in Spain but are owned by investors in the United Arab Emirates.

Question: What are the main differences between the REEM-H and REEM-B robots?

Answer: The REEM-B robot is a prototype of a humanoid robot, and won't be commercially sold. The REEM-H2 robot is designed for the commercial market. It is wheeled and interaction occurs via a touch-screen on the chest.

Question: Wouldn't voice commands be simpler?

Answer: The technology of voice commands has yet to be perfected. Voice commands will eventually be ubiquitous on robots but currently the technology is not sufficiently reliable, accurate, or robust to use.

Question: Can REEM-H2 recharge itself?

Answer: Not yet. But the battery can power the robot for 8 to 10 hours. But we've designed the H2's to have batteries that can quickly and easily be replaced, so we don't see this as a problem.

Question: What tasks can the H2 currently accomplish?

Answer: It can lift objects up to 8 kilograms, so it can transport small packages. It can move up and down ramps, and can navigate indoors using laser scanners. It can be used for video-conferencing, web access, and it can recognize faces and voices.

Question: What tasks can REEM-B accomplish?

Answer: REEM-B was introduced 2 years ago and is targeted primarily at research labs. It is capable of face recognition, it can lift objects up to 14 kilograms, and can walk. It cannot yet climb stairs reliably, or operate outdoors.

Question: How does REEM-B compare to Asimo?

Answer: The Asimo is a beautiful piece of engineering. But Honda has been developing Asimo for almost a quarter century, so they have a lengthy head start. REEM-B can lift heavier objects than Asimo, but unlike Asimo it can't run or dance.

Question: Is there a REEM-C under development?

Answer: Yes, the target is to release REEM-C sometime in 2011. It will be more robust, it will have a greater range of motion, will be able to climb stairs, and will operate up to three times as fast as REEM-B. We also want to reduce the manufacturing costs. REEM-C will have 46 motors, with 16 motors just on the hands, and should be able to operate for 2-3 hours on a battery charge.

Question: Will REEM-C be targeted to consumers?

Answer: No, it will be sold primarily to research labs. Humanoid robots are currently too expensive, difficult to operate, and clumsy to be useful to consumers.

Question: Is software currently the main bottleneck for robotics?

Answer: I think it is. The software is not up the tasks we need it to do, which include reliably responding to voice commands, navigating in the real world, and interacting with objects. The good news is that there are many groups worldwide in the process of improving software programs that can make it possible to perform these tasks. Hardware will not be our major bottleneck once we get to mass production and costs are reduced.

Question: Who is funding all this robotics research?

Answer: We are being directly funded by a company in the United Arab Emirates called PAL Technology. They have been exclusively funding us for the past six years. The Spanish Government is providing a small amount of assistance, the rest comes from Abu Dhabi.

Question: You have been working in robotics for a decade now. What has changed since 2000?

Answer: The pace of robotics development is much faster than it was in 2000. There are more people doing research, the tools are much better, and the funding is considerably greater. We cannot yet fulfill the expectations of customers, but the potential market for autonomous robots is huge.

Question: What technological development will most benefit the robotics industry?

Answer: I would love to see a major advance in robotics vision. Robots employ a number of techniques to navigate, such as 3-d range finders and laser scanners, but cameras are still the optimum technology to use. They are cheap and convey a lot of information, but robots need to do a much better job of interpreting and assimilating that information.

Question: How do you envision the robotics industry in 2020?

Answer: By 2020, robots should be integrated into our life. Robotics will first make an impact in cars that drive themselves. We won't see a robot in every house, but we will see robots starting to become common. But more importantly, robotics companies will be well established, and will have the funds necessary to continue robotics development on their own.

If you liked this article, please give it a quick review on ycombinator, or Reddit, or StumbleUpon. Thanks

Featured articles

Ocean Floor Gold and Copper
   Ocean Floor Mining Company

Thank You
blog comments powered by Disqus