Neural Network Ship Course Control

The automatic instrument of ship maneuvering named automatic steer, it is a important equipment in shipping maneuvering, it's used to maintain a given direction or track. As the increasing of shipping, the need of automatic steer gets greater improvement. The improvement of automatic steer is as same as that of control theories. At 1920's when Sperry of American and Ansuchz of German got essential process in the research of Roland and made the mechanical steer, it's a milestone of shipping maneuvering. Later on the base of development of electronics and the application of them, PID steer appeared which integrated control technology and electronic parts. In the late of 1960's, adaptive theory and computer technology developed speedily, then adaptive steer came forth by using adaptive theory in shipping maneuvering. From 1980's persons began to looking for ways as same as what a driver do, then intelligence steers appeared. They are expert system, fuzzy control and nerve network..This article is discussing the application of nerve network in automatic steer, we can control the direction of ship by a personal window. What I have done are:1) Select the direct inverse nerve network as control strategy. This idea is easy to realize, but with a poor ability, in order to overcome the weakness, we put the interference of environment and beginning starts of appearance into the NNC, so that the network can approach the fact.2) Establish a nonlinear ship model of mathematic. A good ship model can help us great. But we always design control system on the base of linear feedback control theories, it cause almost all the ship models are linear, many details are neglected. As we know that nerve network can approach all kinds of nonlinear system, we can get a precision ship model by talking about characteristic of steer, oar, ship body and the influence of wind, flow and wave.3)Select BP (Back-Propagation) algorithm in nerve network. BP algorithm is in common use and it's one of best algorithms of network. The application of it can improve the dependable. This article discusses the question that how to modify a typical BP neural network to meet application. What mentioned are appearance of...