ON ADAPTIVE TRAJECTORY TRACKING OF A ROBOT MANIPULATOR USING INVERSION OF ITS NEURAL EMULATOR

This paper is concerned with the design of a neuroadaptive trajectory tracking controller. The paper presents a new control scheme based on inversion of a feedforward neural model of a robot arm. The proposed c ontrol scheme requires two modules, The first module consists of an ap propriate feedforward neural model of forward dynamics of the robot ar m that continuously accounts for the changes in the robot dynamics. Th e second module implements an efficient network inversion algorithm th at computes the control action by inverting the neural model, In this paper, a new extended Kalman filter (EKF) based network inversion sche me is proposed. The scheme is evaluated through comparison with two ot her schemes of network inversion: gradient search in input space and L yapunov function approach. Using these three inversion schemes the pro posed controller was implemented for trajectory tracking control of a two-link manipulator. Simulation results in all cases confirm the effi cacy of control input prediction using network inversion. Comparison o f the inversion algorithms in terms of tracking accuracy showed the su perior performance of the EKF based inversion scheme over others.