Neural Network Based Adaptive Feedback Control for Tower Cranes

In this paper, the control problem of 4-DOF tower cranes with friction is studied. In the controller design process, we construct a controller consist of two parts, feedback controller and neural network based adaptive controller. The feedback controller is proposed based on PD control law. Because of mechanical friction, the inputs of tower cranes cannot be applied to the system directly, parts of inputs will be offset. To deal with the friction, adaptive neural network based on radial basis functions are used to approximate friction of tower cranes. In the presence of friction, the controller can transport the payload to specified position with suppressing swing of the payload. We strictly prove the closed loop stability in this paper, using Lyapunov stability theorem. Furthermore, we build the corresponding model in Simulink to verify the performance of controller.