Neural Network-Based Adaptive Learning Control of Nonlinear Crane System

Crane is an important industrial device, and has been widely applied to many fields such as automobile making and marine equipment. This work proposes a neural network-based adaptive learning control scheme for a class of nonlinear cranes with nonstrict-feedback structure and disturbances. Firstly, a prescribed performance function is introduced to limit the swing angle of the load, thereby removing an angle assumption. Then, a series of modified auxiliary variables are designed to simplify the system model. Further, under the framework of backstepping, a learning control algorithm is presented consisting of NN and adaptive laws to ensure that the error approaches zero and other signals are bounded as the iteration goes to infinity. The simulation experiments demonstrate the effectiveness of the proposed scheme.