PSO based backstepping sliding mode controller and observer for robot manipulators

Abstract

This paper examines the structure of an observer-based modified backstepping boundary sliding mode controller (BBSMC) for 2 degrees of freedom robot manipulator. A radial basis functional NN (RBFNN) is combined with BBSMC and adjustment law is derived for the effective control of a robot manipulator. The RBFNN-based observer is used to estimate the tracking position and velocity vector, at the same time RBFNN-based identifier provides an auxiliary torque to the robot manipulator. For both the observer and identifier, a PSO based off-line tuning method is adapted to tune the weights of RBFNN network. The stability of developed controller is checked by utilizing Lyapunov stability function. Finally, the performance of the proposed NNAOBBSMC is validated experimentally.