Performance Optimization of Ship Course Via Artificial Neural Network and Command Filtered Cdm-Backstepping Controller

Abstract

This paper proposes a robust nonlinear ship course controller, under the control of which the system is globally asymptotically stabilized with high control quality. The proposed controller is synthesized by combining coefficient diagram method and command filtered backstepping based on first order filter to avoid the complex analytic derivation of the virtual control, the controller parameter are tuned using radial basis function neural network, It can not only obtain a higher accuracy in ship course controlling, but also infinitely approach the nonlinear system with quicker and more stable convergence. The simulation results illustrate that the projected controller shortens the settling time evidently with good system stability. It has a better performance than the traditional controllers.