SGD-based adaptive NN backstepping control design for a linear motor system

In this paper, the motion control problem for a linear motor system is solved by using a new adaptive neural network (NN) control strategy. This control method combines the stochastic gradient descent (SGD) algorithm and the adaptive NN backstepping technique, which results in good control performance in dealing with system uncertainties. The weights, widths, and centers of NN can be continuously updated and optimized such that the system uncertainties can be well approximated by the NN. Based on Lyapunov stability theory, it can be proven that the motion error eventually converges to a small neighborhood of the origin. Finally, the practicality of this method is verified through experiments.