Internal model control using neural network for ship roll stabilization

Abstract

In this paper, a neural network (NN) based on internal model control (IMC) is developed to adjust control parameters for roll motions of a container ship. Controller architecture, which combines neural network with internal model control, has been outlined and its effectiveness demonstrated on the container ship roll stabilizer. The control signal error is used with back-propagation algorithm to update the weights of the neural controller. In conclusion, the neural network based on internal model control systems are analyzed, and compared to classical PID control results. As can be seen from numerical results, the NN based on IMC is implemented successfully to reduce roll amplitude.