A trajectory generation method for biped walking based on neural oscillators

Abstract

A new trajectory generation method is proposed based on two neural networks. One oscillatory network is designed to generate foot trajectory, and another set of neural oscillators can generate the center of mass (CoM) trajectory in real-time. Using a motion engine, the characteristics of the workspace is mapped to joint space. Sensory feedback is applied to modulate the generated trajectories in real time to improve walking adaptability and stability. The developed control strategy is tested using a humanoid robot on slope terrain. The results show that the robot can successfully walk on terrains with varying slopes through autonomous adjustment of its walking patterns.