Abstract
The proposed PD-type position control law for manipulator robots adjusts the proportional gains based on the desired position. This capability improves the PD controller’s efficiency in trajectory tracking in terms of accumulated position error and energy efficiency, even when the manipulator’s dynamic model parameters are unknown. To accomplish this goal, a Radial Basis Function interpolation network, trained offline to avoid higher computational demands, replaces each proportional gain. The Lyapunov method ensures the system’s stability, and its effectiveness in position control is further assessed under parametric uncertainties and external perturbations through Monte Carlo analysis and Kruskal-Wallis statistical tests. Matlab simulations on a two-degree-of-freedom manipulator arm following an owl-shaped trajectory demonstrate that, in trajectory tracking, the proposed controller achieves improved ℒ2 norm and energy efficiency compared with the PD controllers of Takegaki-Arimoto and Tanh(•) with bounded actions.