Evolutionary Learning Acquisition of Optimal Joint Angle Trajectories of Flexible Robot Arm

This paper proposes the evolutionary learning acquisition method of the optimal joint angle trajectories of a flexible robot arm using the genetic algorithm is proposed, and the effects of the optimal joint angle trajectories obtained by the present evolutionary learning acquisition method on the residual vibration reduction are ascertained numerically and experimentally. In the construction of the evolutionary learning acquisition algorithm of the optimal joint angle trajectories, the joint angular velocity curves are depicted with fifth-order polynomials, and, by considering the boundary and constraint conditions, they are expressed by four parameters. Then, the residual vibrations of the flexible robot arm are expressed as a function of the chromosome consisting of four parameters, namely, four genes, and a fitness function of the genetic algorithm for the residual vibration reduction is defined. Furthermore, the numerical calculations have been carried out, and it is confirmed that the residual vibrations almost disappear. Moreover, the experimental results are demonstrated, and the usefulness of the present evolutionary learning acquisition method of the optimal joint angle trajectories of the flexible robot arm using the genetic algorithm is ascertained experimentally.