Hybrid neuro-fuzzy control approach of robot manipulators

Abstract

Recently, robot manipulators are expected to perform more sophisticated tasks. Thus, these manipulators have to be intelligent enough to work in an unknown and constrained environment. In this paper, we are interested by the uses of a hybrid neuro-fuzzy control approaches for robots manipulators moving in such environment. For this purpose, two adaptive neuro-fuzzy control structures are presented, an external effort control structure and an hybrid force/position control computing a Reinforcement Model (RM). Motivated by the hybrid force/position control problem, a systematic design procedure for fuzzy rules generation and optimization is proposed. The proposed neuro-fuzzy controller (NFC) is constructed with respect to a specific methodology. A 2 DOF planar manipulator simulation is presented and results discussed.