A Closed-Loop Central Pattern Generator Approach to Control NAO Humanoid Robots’ Walking

Biped locomotion has attracted much attention in recent years. The most successful implemented methods in this area are based on two approaches, Central Pattern Generator (CPG) and Zero Moment Point (ZMP). In CPG based controllers, it is difficult to figure out how the feedback should be introduced to the CPG network. In this study, we introduce a closed-loop CPG-based controller which improves stability of the robot, using its center of pressure (COP) as a real time feedback control. It is shown that by using the deviation of real COP relative to the reference COP, the robot will be able to face with unexpected disturbance or sudden external forces of environment. Another problem of CPG based methods is their huge number of parameters. We challenge this problem by incorporating Genetic Algorithm (GA). Furthermore, performance of the proposed control system is verified on NAO humanoid robot, using RoboCup 3D soccer simulation server