LEARNING HYBRID POSITION FORCE CONTROL OF A QUADRUPED WALKING MACHINEUSING A CMAC NEURAL-NETWORK

Learning control algorithms based on the cerebellar model articulation controller (CMAC) have been successfully applied to control non-linea r robotic systems in the past. Most of these previous works are focuse d on the position controls of manipulators. In this article, a CMAC-ba sed learning control method for the hybrid force/position control of a quadruped walking machine on soft terrains is presented. The relation ship between the foot force and the control variables is derived for v arious force control methods. By using the CMAC to approximate the dyn amics of one leg, we are able to demonstrate the improved control accu racy without the exact leg model. The same concept is extended to the control of a quadruped walking machine. (C) 1997 John Wiley & Sons, In c.