Optimal Trajectory Planning for Robot

Abstract

A gait control system is designed for walking a slope movement according to the features of biped robot, Two BP networks are introduced into train the walking trajectory of robot and the optimal trajectory is generated by the new evolutionary approach based on particle swarm optimization. Additionally, online environmental information is collected as neural network input, and necessary joint trajectory is output for the movement. Simulations with Matlab and experiments on NAO humanoid robot testify the efficiency of the method.