Fuzzy Control of Vertical Jumping With a Planar Biped

Abstract

This paper develops a control strategy to produce vertical jumps in a planar biped robot as a preliminary investigation into dynamic maneuvers. The control strategy was broken into two functional levels to separately solve the problems of coordination and execution of the jump maneuver. A high-level fuzzy controller addresses the complexities that arise from the system’s hybrid nonlinear dynamics and series-elastic actuators embedded in the articulated legs. A novel fuzzy training scheme is used because the system is too complex for traditional training methods. A low-level controller is based on a state machine that sequences the legs through the phases of a jump. The modular nature of the control strategy allows quick adaptation to other dynamic maneuvers. Validity is demonstrated through dynamic simulation and testing with the experimental biped KURMET which result in stable successive jumps over a range of heights.