A Learning Behaviour Based Controller for Maintaining Balance in Robotic Locomotion

A new approach to robotic balance control under disturbance, called the Behavior Based Locomotion Controller (BBLC), is presented. This architecture implements a Behavior Based Control (BBC) architecture to generate new balancing strategies to compensate for unknown disturbances in the environment. This is achieved by dividing bipedal locomotion into several task-space motions, defining multiple balancing behaviors for each task-space motion and using a reinforcement learning algorithm to determine which behavior combinations result in new balancing strategies. The controller is implemented on ABL-BI, a 13 degree of freedom bipedal robot. Three disturbances cases are examined: a push disturbance, a step under one foot and a slope. For each case, the BBLC is able to generate a new balancing strategy that increases the robustness of the system to the disturbance. Additionally, an evaluation of the selected balancing behaviors is completed using a stability analysis of the linear inverted pendulum.