The adaptive sliding mode control based on a fuzzy neural network for manipulators

Abstract

In this paper, a new control scheme for the computer control of a robot manipulator is provided, where a fuzzy neural network is used to identify the robot dynamics, a feedforward central algorithm is designed to compensate for variations in mass, gravity, Coriolis and centrifugal terms, and the discrete sliding mode control is applied to reduce further the remaining system uncertainty for precise position tracking. The whole control algorithms are developed in a discrete-time form which is very suitable for the computer control of robot manipulators.