An adaptive controller with an orthogonal neural network and a third order sliding mode observer for robot manipulators

In this paper, An adaptive controller with an orthogonal neural network (ONN) and a third order sliding mode(TOSM) observer for robot manipulators is proposed. Firstly, the TOSM observer is designed to observe joint velocities. Then, the ONN is designed to compensate robot dynamic uncertainties on line inside a computed torque control structure. Therefore, the proposed controller allows only position measurements due to the TOSM observer and achieve highly accurate trajectory tracking performance due to the ONN’s uncertainty compensation. Finally, computer simulation for a 2-DOF manipulator is performed to show verify the effectiveness of the proposed controller.