Adaptive neural network control for ship steering system using filtered backstepping design

作者：Ren Junsheng^*; Liu Lu

来源：Journal of Applied Sciences, 2013, 13(10): 1691-1697.

DOI：10.3923/jas.2013.1691.1697

• Adaptive Control
• Adaptive controllers
• Adaptive neural network control
• Back-stepping method
• Backstepping controller
• Radial basis function neural networks
• Ship steering
• Time varying signal

As regards ship course control, the ship is characterized by a nonlinear function with uncertainties, representing maneuvering characteristics. This study addresses the design of adaptive controller for ship steering system. The control objective is to drive the course to track a prescribed time-varying signal. We use filtered backstepping method to design the control law. Radial Basis RBF) neural network learns the system's uncertainties and nonlinearities online. An adaptive law is combined with a control design including a filtered backstepping controller and RBF neural network approximator. Our analysis revealed even if there is no a priori knowledge about ship's system dynamics, the design can guarantee the ultimately uniformly boundedness for ship steering closed-loop system. Furthermore, the controller contains only one online learning parameter and the laborious differential computation in conventional ship method become unnecessary. Ship maneuvering scenario is simulated to verify the effectiveness of our approach.

• 出版日期2013