Study On Fuzzy And Neural Network Control System For Ship Steering

Along with the development of shipping business, ships are becoming bigger, faster and more intelligent. Because there is prodigious inertia when a bigger ship is sailing, the high performance of maneuver is demanded. On the other hand, with the increase of ship number and the appearance of special ships, the sea-lane and the port are becoming relative narrow. It is becoming more complicated and more difficult. To ensure safety and raise economic it is necessary to adopt the new control theory, control strategy and technique. Fuzzy control, neural networks control and genetic algorithm in ship steering motion control system are systemically investigated in this thesis. This thesis concentrates on studying no-model self-study and adaptive ship course control system by using fuzzy logic, neural network and genetic algorithm.Based on the characteristics of ship steering the membership functions of input and output lingual variables and control rules are formed in this thesis. So a kind of course fuzzy controller is designed. Aimed at overcoming the shortage of basic fuzzy controller, a kind of two stage fuzzy controller with scaling factors self-regulating and fuzzy PI controller is presented. When it is used in course control system, system' s anti-disturbance and response speed are satisfied. By establishing fuzzy controller' s describing function and analysis of system' s stability, the ship course fuzzy control system is stable.Adaptability is the target for fuzzy control system. Two kinds of ship steering motion adaptive fuzzy control system are proposed. One is model reference fuzzy adaptive control system, in which the fuzzy logic inference is used in adaptive algorithm. So it solves the problem of complicated adaptive algorithm in conventional model reference adaptive control system. It raises system's stability. The other is ship course self-studying fuzzy control system, which is based on the learning mechanism of robot repetition control. The thesis analyses the system' s characteristic.After analyzing the feed ford neural networks structure, algorithm and astringency, which are used very popular in control field, the thesis proposes that the ship steering dynamical characteristics is learnt with RBF neural network. The simulation results show that the RBF neural network has faster learning speed and higher prediction precision than BP neural network. In the respect of neural networks control for non-linear and uncertain system, a review of some available control strategy is made. Combining neural networks control and conventional control strategy Supervised Learning, No Supervised Learning and Reinforcement Learning neural networks self-studying and adaptive control systems for ship course control areproposed. The thesis studies particularly their characteristics.Although fuzzy logic system can express knowledge easily and has stronger fuzzy logic inference ability, it is difficult in self-studying respect. Artificial neural networks have particular superiority in self-studying and function approximation respect. Based on the integration of fuzzy logic system and neural networks the thesis constructs a fuzzy neural network controller, which is used in ship course self-studying adaptive control system. Because the fuzzy logic system is realized in neural network, the fuzzy neural network controller' s self-studying ability is raised greatly. Substituting improved genetic algorithm for back propagation algorithm in fuzzy neural network controller ship course fuzzy neural network control system' s real-time ability is raised. According to ship course control' s feature and economics the fitness function is designed in genetic algorithm. These works laid a function for practical applications to ship steering motion control.From the engineering realization, a kind of ship course and trace intelligent control system scheme is proposed. The thesis points out that the expert decision-making mechanism, which can decide control strategies and control means in real-time, is needed in it. At the s...