Locomotion Control of a Hexapod Robot Based on Central Pattern Generator Network

摘要

This paper presents locomotion control of a hexapod robot based on central pattern generator (CPG) network. To do this task, the followings are done. First, kinematic modeling of one leg with four links and four rotational joints of hexapod robot is proposed using Denavit Hartenberg (DH) convention. Second, the walking gait of hexapod robot is generated by central pattern generator network. Third, mapping functions are proposed to map the output of CPG network into a desired trajectory of the end effector of the hexapod robot. It is difficult and complex to get the angles of four rotational joints of one leg using inverse kinematics. To solve this problem, differential kinematics algorithm is used to realize the end effector following the trajectory obtained by the mapping functions. Finally, simulation and experimental results for walking motion of one leg of the hexapod robot are shown to prove the effectiveness and applicability of the proposed controller.