Pages

October 14, 2009

David Calkins Interview by Sander Olson

David Calkins is a roboticist who is the President of the Robotics Society of America. He has taught robotics at San Fransisco State University, designed and built hundreds of robots and has worked with NASA on the robotics education project. He recently created a company to produce robots for the home.

There are two embedded videos below. The videos are about
* David Calkins on Humanoid Robots
* David Calkins reveals the basics of putting together and programming your very own android

Question 1: How much progress has been made in robotics during the past decade?

Answer: Quite a bit of progress has been made during the past ten years. Motor technology has improved, as have feedback loops. Robotic tasks that used to be impractical, such as walking, are now not only feasible but are available for under $1,000.



Video: David Calkins on Humanoid Robots


Question 2: How long have you been in the field? How many robots have you constructed?

Answer: I have been in this field for 11 years. I have constructed hundreds of robots, mostly 14 inch models that are humanoid and can run, play soccer, and do cartwheels. Many people who build robots become obsessive about it. For me, I enjoy building the robots as much as having them.


Question 3: How long will it be before a trillion dollar robotics industry emerges?

Answer: If one includes the consumer industry, the military, and the manufacturing industry, it may already be approaching a trillion dollar industry. But if we don't currently have a trillion dollar robotics industry than we soon will.

Question 4: How important a role is the military playing in robotics development? Are they the leading driver of robotics technology?

Answer: No, the robotics field is being pushed forward from multiple sources. Military robots, house cleaning robots, factory robots - they all have different development tracks. The DOD is funding many robotics companies, most of which are producing noncombat models. These companies can use the profits generated from military contracts to invest in civilian robotics technology, so there is a positive spinoff.

Question 5: Tell us about the robogames competition.

Answer: The aim of the robogames competition is to get robot builders from all over the world to congregate and cross-pollinate. We want to have the best and brightest robotics experts from all over the world, to share their technologies, and to discuss how to build better robots.


Question 6: To what extent is a lack of standards hampering robotic technology?

Answer: Certain standards for robotics already exist. For instance, the servos in robotics are all standardized. Unlike computers, robots don't necessarily require operating systems. I could build a walking humanoid robot without an operating system. At some point in the future, a set of robotics technologies will need to emerge, but it isn't impeding robotics development at this early stage.

Video: David Calkins reveals the basics of putting together and programming your very own android


Question 7: What industry will be the biggest user of robots during the next decade?

Answer: Military and law-enforcement will be major consumers simply because of their large budgets and their desire to protect their personnel from direct harm. As robots go from simple, stupid devices that vacuum floors to more sophisticated versions that can differentiate objects in their environment, the programming challenge becomes exponentially harder. So many technological hurdles have to be surmounted in order for robots to operate effectively in an open, unstructured environment that consumer markets will not be a big user of computers in the next decade.


Question 8: Why is it so difficult to have a robot operate in an unstructured environment?

Answer: Human development has 50,000 years of evolution going for it. Humans are adept at noticing subtle differences, such as dog hair on a carpet, or the difference between a print and a wine stain on a rug. It is quite difficult to program a robot to detect these differences. A straight set of rules won't work, because there are too many exceptions to the rules. Developing that level of artificial intelligence will take decades.


Question 9: When will robotic cars become commonplace?

Answer: We won't see that industry take off until robots can distinguish between, say, an empty plastic bag and a basketball. Without that ability, robotic cars would stop at every impediment in the road and could actually cause crashes. So a robotic car will require a vision system with sophisticated pattern-recognition systems.


Question 10: How serious are power supply problems to the development of robotics?

Answer: The power-supply problem has essentially been overcome, thanks to the mobile electronics industry. So much effort has been put into batteries for power tools, cellphones, and laptops that the power-supply problem has almost been eliminated.


Question 11: How necessary is sentience for robotics development?

Answer: Sentience is unnecessary. Most robots will be single-purpose devices, and won't require sentience to effectively do their job.


Question 12: The roboticist Hans Moravec has argued that computing power is key to robotics development. Do you agree?

Answer: It is absolutely necessary for sophisticated robots. A simple vacuum cleaning robot requires virtually no computing power, and could have existed 20 years ago. But a versatile robotic butler capable of fixing meals, shopping, etc., would probably require a huge amount of processing power.


Question 13: How much progress do you see robotics making in 2010-2020 decade?

Answer: Robotics development is definitely following a Moore's law curve. We will probably see 10 times as much progress in robotics technology in the next decade as in the last ten. The field is clearly expanding exponentially, and it is only a matter of time before robots become ubiquitous in society.

blog comments powered by Disqus