A GENETIC ALGORITHM FOR TORQUE OPTIMIZATION OF AUTONOMOUS VEHICLES IN HIGH INCLINED TERRAINS

DIEGO  G. G. ROSA; MARCOS  ANTONIO MEGGIOLARO; MARCO  AURÉLIO CAVALCANTI PACHECO

doi:10.20906/CBA2022/445

DIEGO G. G. ROSA Pontifícia Universidade Católica do Rio de Janeiro
MARCOS ANTONIO MEGGIOLARO Pontifícia Universidade Católica do Rio de Janeiro
MARCO AURÉLIO CAVALCANTI PACHECO Pontifícia Universidade Católica do Rio de Janeiro

DOI: https://doi.org/10.20906/CBA2022/445

Keywords: Genetic algorithms, Torque optimization, Stability analysis, Mobile robotics

Abstract

In this paper, a strategy based on genetic algorithms for mechanical torque optimization is implemented. The work is based on an electric 4WD (four-wheel drive) vehicle. This configuration is interesting as it can be generalized for mobile robots, off-road vehicles or even electrical wheelchairs. A multi-objective criterion is used. In usual applications, high velocities are desired to the trajectory accomplishment and, simultaneously, a stable movement have to be executed, even in terrains with high inclinations. Hence, to guarantee the stability in high inclined plans is the main challenge in this work. The fitness function here used consists in evaluate torques for a simplified dynamic model, while electrical, geometrical and mechanical vehicle character-istics are restricted. The results show interesting information about the systems behaviour, as the maximum inclinations that the vehicle can rise and the possible torques that can be executed to accomplish a trajectory in a maximum velocity. Simulations of a theoretical model and experimental tests developed in a mobile robot assure the results.