Biped Landing Pattern Modification Method and Walking Experiments in Outdoor Environment

Abstract

Many researchers have studied walking stability control for biped robots, most of which involve highly precise acceleration controls based on robot model mechanics. Modeling error, however, makes the control algorithms used difficult to apply to biped walking robots intended to transport human users. The “landing pattern modification method” we propose is based on nonlinear admittance control. Theoretical compliance displacement calculated from walking patterns is compared to actual compliance displacement, when a robot's foot contacts slightly uneven terrain. Terrain height is detected and the preset walking pattern is modified accordingly. The new biped foot we also propose forms larger support polygons on uneven terrain than conventional biped foot systems do. Combining our new modification method and foot, a human-carrying biped robot can traverse uneven terrain, as confirmed in walking experiments.