Iterative learning control for nonlinear systems based on neural networks

An error-backpropagation neural network (NN) is applied to iterative learning control for a class of nonlinear control systems. It realizes full-state feedback control for nonlinear systems via iteration. It avoids the demands of traditional PID learning control due to the generalizability of the neural network. Meanwhile, it avoids the difficulties of online control of fast systems. The gradient-type learning control algorithm is derived, which does not strictly depend on the model of the controlled system. Simulation results show that the new scheme is efficient for large unknown nonlinearity.