Intelligent control of the running stride of a simulated jointed leg

This paper presents an intelligent controller for a simulated human-like jointed leg during running. A neuro-fuzzy takeoff controller propels the leg off the ground and produces a running stride of a specified size. A neuro-classical ballistic controller moves the airborne leg along a specified trajectory, scaling any obstacles. A similar landing controller returns the leg to the takeoff position. High accuracy is achieved on a wide range of stride sizes.