Dynamic neuro-fuzzy adaptive control for flexible-link manipulators based on dynamic inversion

Abstract

A dynamic neuro-fuzzy (DNF) adaptive control system is presented in this paper for the trajectory tracking of a flexible-link manipulator with poorly known dynamics, where the robot tip vibration is measured by position sensitive detectors (PSDs). Based on the singular perturbation method and two time-scale decompositions, the robot dynamics is approximated by a slow subsystem of an equivalent rigid arm and a fast subsystem of flexible mode. Then, the DNF adaptive controller based on dynamic inversion is designed for the tracking control of the equivalent rigid arm, while a fuzzy proportional derivative (PD) type controller is used to stabilize the elastic dynamics using PSDs measurement feedback. The NF learning algorithm and the stability proof of the closed-loop system are given, and an upper bound for the singular perturbation parameter is also obtained. Finally, the real-time experiment is conducted to show the viability and effectiveness of the proposed control approach.