Neural Network MLP with Sliding Mode Controller for Robotic Manipulator

In this paper, MLP network with sliding-mode controller and neural control is designed to the joint position control of two-link robot manipulators for periodic motion and predefined trajectory tracking control. Because the nonlinear nature robot, first the nonlinear and robust method sliding mode as a control theory for the system is chosen. Chattering phenomenon is always a problem of sliding mode controller is implemented. To reduce this effect, neural network in combination with a sliding mode controller is used, to estimate part discontinuous of sliding mode controller. In the next step, the robot arm system by two other neural networks was controlled as intelligent control. Simulation results show that this method is very effective in reducing Chattering. Finally, a robustness analysis was also performed and resistance of intelligent control methods is considered. KEY WORD: MLP; Neural network; Nero control; Robotic manipulator; Sliding mode.