Locomotion control of quadruped robot based on central pattern generators

Abstract

The Central Pattern Generators (CPGs) have gained the favor of many researchers due to their stable limit circle properties in recent years. However, simple CPGs can merely implement open-loop control, which means attitude of robot body cannot be controlled effectively. In this paper, we present a controller for quadruped robot locomotion in the flat and slope. The attitude reflex is implemented as sensory feedback augmenting the CPGs. The control strategy is verified on a 3D quadruped robot platform in SimMechanics. And finally we realize the stable straight-line walking for the quadruped robot in the flat terrain and the slope.