Quadrupedal Robots’ Gaits Identification via Contact Forces Optimization

Abstract

The purpose of the present paper is the identification of optimal trajectories of quadruped robots through genetic algorithms. The method is based on the identification of the optimal time history of forces and torques exchanged between the ground and the body, without any constraints on leg kinematics. The solutions show how it is possible to obtain similar trajectories to those of a horse’s walk but obtaining better performance in terms of energy cost. Finally, a map of the optimal gaits found according to the different speeds is presented, identifying the transition threshold between the walk and the trot as a function of the total energy spent.