Walking Pattern Design and Balance Control of a Quadruped Platform

Abstract

This paper addresses the design and generation of walking patterns for a quadruped platform as well as a balance control for it. Using Multi Objective Genetic Algorithm, two optimized gaits are generated and implemented in the real platform as well as in a simulation environment created in V-Rep. Moreover, a balance control is derived, based on a kinematics approach and an empirical methodology. To obtain a consistent response to disturbances and maintain the stability of the robot, the balance control is designed proportionally to the speeds obtained by the embedded accelerometer measurements. Finally, the results obtained on the real platform for both the gaits generated and the balance control in response to external disturbances are described.