Multi-Objective Design/Control Optimization on the Power Train of Robot Manipulators using a Genetic Algorithm

This paper deals with a multi-objective design and control optimization on the drive train for robot manipulators. The objective functions to be minimized are: i) the total weight of the components installed on the arm, finding the best commercially available motor/gearbox combination in a catalog list, ii) the tracking error for a desired trajectory and iii) the motor energy consumption. To simultaneously solve the mechanical and electrical dynamics of the drive train, a corrective composite control was proposed, then design and control parameters are involved in the system at the same time. This led us to formulate a concurrent optimization problem with continuous and discrete variables, where selection criteria of components on the drive train, dynamic operation requirements, changes at inertial parameters and the control gain tuning are used to ensure feasible solutions thus achieving good performance of the whole system. The second generation Multi-Objective-Genetic Algorithm method, called Non-dominated Sorting Genetic Algorithm-II (NSGA-II), is used to solve the optimization problem. The method is proved redesigning a planar three dof robot manipulator by selecting a particular solution of the obtained form called Pareto front for a required task.