Evolutionary optimal trajectory planning for industrial robot with payload constraints

This paper presents a new general methodology based on the evolutionary algorithms—elitist non-dominated sorting genetic algorithm (NSGA-II) and differential evolution (DE)—for optimal trajectory planning of an industrial robot manipulator (PUMA560) by considering payload constraints. The aim is to minimize a multicriterion cost function with actuator constraints, joint limits, and payload constraints by considering dynamic equations of motion. Trajectories are defined by B-spline functions. This is a nonlinear constrained optimisation problem with five objective functions, 32 constraints, and 252 variables. The multicriterion cost function is a weighted balance of transfer time, total energy involved in the motion, singularity avoidance, joint jerks, and joint accelerations. A numerical example is presented for showing the efficiency of the proposed procedure. Also, the results obtained from NSGA-II and DE techniques are compared and analysed. A comprehensive user-friendly general-purpose software package has been developed using VC++ to obtain optimal solutions using the proposed DE algorithm.