Dynamic self-tuning PD control for tracking of robot manipulators

Abstract

Motivated by the passivity-based robot control scheme under unknown dynamics, a self-tuning PD controller with bounded time-varying gains is proposed for trajectory tracking of robot manipulators. The tuning procedure is synthesized on-line by means of a neuro-fuzzy recurrent scheme which produces smooth control input for smooth desired trajectories. Semi-global exponential stability of the closed-loop system is guaranteed via Lyapunov theory. Simulation results make evident and viable the potential characteristics of the proposed controller.