The Application of Human Motor Control Theory to RoboticsIntroductionIn this article I will examine how some of the theories about skilled physical movements in humans might be applied in the field of robotics. These theories can be adapted to any type of robotic control systems but I will focus primarily on their practical use in analog control systems. The experienced amateur working in the field known BEAM robotics should have no trouble experimenting with the circuits presented here. These ideas are quite different from the usual approach to complex robotic movements. In conventional robots extensive calculations are performed to plan each movement. A complicated movement may be broken up into several small movements with lengthy pauses between them as the processor plans the next move. The approach I suggest here does not require any calculation during the movement. Instead the robot carries out a selection process and then performs a complete, complex motion using pre-established parameters. The article is divided into the following sections.
Every field of study has its own unique terminology. Bringing together the fields of psychology, control theory, and robotics creates special terminology problems. Of particular note:
AcknowledgementsThe idea to use robotic models of human motor programs has followed an obscure but seemingly inevitable course. My preliminary ideas on complex control systems for robots go back to 1986 and Marvin Minsky's wonderful psychology book "The Society of Mind." My shorthand constructions, F-lines, I-lines, and P-lines, are derived from this book although they serve quite a different purpose. Ben Hitchcock pointed out that my the hierarchical switch, a hardware approximation of some of Minsky's concepts, resembled subsumption architecture. I was frankly doubtful, as I had previously seen a totally different kind of structure that claimed to use that subsumption architecture. In the summer of 2001 I finally got around to investigating this and found Rodney Brooks' publications at MIT. While confirming that my control system did, in fact, meet the minimum requirements of subsumption architecture I also stumbled across his description of "Squirt", a tiny robot that hid in the dark until it heard a noise. That prompted me to design my analog robot Hider with better survival capabilities than Squirt. I began to design Hider's control circuit using my early methods but quickly discovered I was using a brick to kill a mosquito. Hider's behaviour is sequential, not hierarchical, and this led to the obvious solution of using a chain of Nv neurons similar to the ones used on Mark Tilden's prototype walker. During a discussion with David Perry about using neuron chains for higher level control I suddenly made the connection with my 1991 Kinesiology courses which explored motor programs and closed-loop theories of human movement. The result is this article.Mark Tilden has been very much an inspiration and is largely responsible for my drive to do more with analog control circuits. His basic circuit element, the Pulse Delay Circuit or nervous neuron, forms the basis of my simulated motor program. Wilf Rigter has been a true mentor in the field of electronics. He clarified my earliest questions on the inner workings of the nervous neuron and also gave me the basic ideas that are incorporated into my Pulse Width Modulation circuit. Robey R. Reed inadvertently got me interested in Kinesiology. More important, he and his instructors taught me how to move. This more than anything else has convinced me that Motor Programs of some sort are used to produce skilled human movements. Finally, a tip of the hat to Dave Simmons. His innocent question gave me a sudden insight into how to implement simple deceleration.
Special thanks to Clifford Boerema, Dave Hrynkiw, David Perry, and Wilf Rigter for
reviewing my first draft. Their suggestions have made this a much better article.
ReferencesAdams, Jack A.; Issues for a closed loop theory of motor learning. In G.E. Stelmach (Ed.), Motor control: Issues and trends, New York, Academic Press, 1976. Brooks, Rodney A. and Anita M. Flynn; Fast, cheap and out of control: A robot invasion of the Solar System, Journal of The British Interplanetary Society, Vol 42, pp 478-485, 1989. Fox, Edward L., Richard W. Bowers, and Merle L. Foss; The Physiological Basis of Physical Education and Athletics, 4th Edition, Iowa, William C. Brown, 1989. McArdle, William D., Fank I. Katch, and Victor L. Katch; Exercise Physiology, 2nd Edition, Philadelphia, Lea & Febiger, 1985. Minsky, Marvin; The Society of Mind, New York, Simon and Schuster, 1986. Nakayama, Masatoshi;Dynamic Karate, Tokyo, Kodansha International, 1966. Schmidt, Richard A.; Motor Skills, New York, Harper & Row, 1975.
|