基于神经网络预报的专家控制器设计

An Expert Controller Based on Neural Network Predication

采用人工神经网络对船舶操纵动态特性进行在线辨识,并提出了一种基于神经网络预报的专家控制算法用于船舶自动操舵仪的设计。该算法在保证控制系统稳定的前提下,能实时跟踪船舶航行环境的变化和船舶操纵动态特性的变化,逐步提高自动舵的控制性能。仿真结果表明了新算法的有效性,另外,算法简单,易于在线实现。

The on-line identification of a ship's dynamic characteristic with neural network is adopted. This paper proposes an expert control algorithm with which a ship's autopilot is designed. This algorithm is based on the predication of ship's motion by neural network. Using this algorithm,the control system is stable,and real-time track for ship's navigation conditions and the variations of dynamic characteristic is realized. Simmlation result shows that the effectiveness is good and due to its simplicity,it can be as an on-line application.