Central Pattern Generators for Gait Generation in Bipedal Robots

Abstract

An obvious problem confronting humanoid robotics isthe generation of stable and efficient gaits. Whereaswheeled robots normally are statically balanced andremain upright regardless of the torques applied to thewheels, a bipedal robot must be actively balanced,particularly if it is to execute a human-like, dynamicgait. The success of gait generation methods based onclassical control theory, such as the zero-moment point(ZMP) method (Takanishi et al., 1985), relies on thecalculation of reference trajectories for the robot tofollow. In the ZMP method, control torques aregenerated in order to keep the zero-moment point withinthe convex hull of the support area defined by thefeet. When the robot is moving in a well-knownenvironment, the ZMP method certainly works well.However, when the robot finds itself in a dynamicallychanging real-world environment, it will encounterunexpected situations that cannot be accounted for inadvance. Hence, reference trajectories can rarely bespecified under such circumstances. In order to addressthis problem, alternative, biologically inspiredcontrol methods have been proposed, which do notrequire the specification of reference trajectories.The aim of this chapter is to describe one such method,based on central pattern generators (CPGs), for controlof bipedal robots.