Active Disturbance Rejection Control for humanoid stable walking

Abstract

This paper presents an application of the active disturbance rejection control approach for humanoid robot stable walking, which is a control strategy based on the extended state observer and cascade integrator canonical form features from modern control theory to deal with disturbances, nonlinearities and unmodeled dynamics. The proposed controller is based on the 3D linear inverted pendulum walking pattern generation, and is designed to be triggered only if disturbance signals are detected. This control technique was tested on the model of a humanoid robot using numerical simulation software, showing successful results that motivate to implement it in a real humanoid robot platform.