Multi-objective Optimization Based Self Tuning Robot Manipulator Controller

The multi-joint robot manipulator has nonlinear and strong coupling characteristics. It's motion planning can be obtained by multi-objective optimization control with minimum torque motion and position deviation. In this paper, NSGA-II and preference information based RV-NSGA-II multi-objective optimization algorithm are used to optimize the control of two link manipulator joints respectively. On the basis of the Kinetic model of two-link manipulator, the proportional, integral and differential coefficients of the PID controller can be online adjusted by the multi-objective optimization algorithm for the desired motion planning path, so that the motion of two link manipulator can be controlled. The optimization results show that the self-tuning PID controller of NSGA-II can effectively control the joints according to the established target, especially the proposed preference information based RV-NSGA-II multi-objective algorithm has less deviation on joint position control.