Parallel-populations genetic algorithm for the optimization of cubic polynomial joint trajectories for industrial robots

Abstract

In this paper a parallel-populations genetic algorithm procedure is presented for the obtainment of minimum-time trajectories for industrial robots. This algorithm is fed in first place by a sequence of configurations then cubic spline functions are used for the construction of joint trajectories for industrial robots. The algorithm is subjected to two types of constraints: (1) Physical constraints on joint velocities, accelerations, and jerk. (2) Dynamic constraints on torque, power, and energy. Comparison examples are used to evaluate the method with different combinations of crossover and mutation. © 2011 Springer-Verlag.