Towards Dynamic Quadruped Locomotion: Development of a CPG-driven Foot Trajectory Generator

Abstract

Legged locomotion has been a widespread daily activity of animals for millions of years, and many of its subtleties are founded on imposed conditions for survival. Agile movement in environments built for human locomotion could be achieved by extracting and implementing key aspects of legged locomotion in artificial creations. This work proposes a novel high-level control system module for dynamic quadruped robot locomotion, integrating the rhythmic developing capabilities of CPGs with foot trajectory generation based on Bézier curves. The proposed system utilizes CPG output signals as driving parameters for a foot trajectory generator based on Bézier curves and is built using a Matlab/Simulink simulated environment, with tests being carried out to validate its quadruped locomotion aptitude with regards to limit-cycle convergence and establishment of synchronized and stable phase shifts corresponding to desired gaits. The results point toward a sound performance of the proposed strategies, showing great promise regarding its locomotion capabilities. Lastly, additions to enhance the system's effectiveness and enable its use in a complete locomotion control system are suggested to be further explored in future works.