Bipedal Robot Locomotion on a Terrain with Pitfalls

In this paper a locomotion control system for bipedal robot is proposed to provide desirable walking on a terrain and skipping over a pitfall preventing the robot from falling in it. The proposed strategy is a combination of motion optimization based on particle swarm optimization algorithm and utilization of mode switching at the higher level controller. The model for bipedal robot is a compass gait model but the presented method is general and could be appropriately extended and generalized for other complicated models. Principles of minimalistic designs are also respected and simple central pattern generator and simple mechanical feedback control are used to produce and maintain desirable motion patterns of the robot.