Neural Network Adaptive Control of a Robotic Manipulator with Dynamic Characteristic and Unknown Disturbance

Robotic manipulator is widely used for its unique advantages such as good human-machine interaction, strong flexibility, and remote controllability. However, complex dynamic characteristic and unknown disturbance make the controller design difficult for robotic manipulator. To overcome the above obstacle, neural network (NN) based adaptive controller is presented, where NN is used to approximate the unmodeled dynamic characteristic and unknown disturbance. The approximation error is addressed by robust control law with adaptive gains. In addition, the tracking performance and asymptotic stability of the closed-loop system is proved using the Lyapunov theory. Finally, comparative simulations of a two-DOF manipulator in different conditions are conducted to validate the effectiveness of the proposed controller.