RBF Neural Network Fuzzy Control For Ship Course Keeping And Integrated Monitoring Data Communication System

There are three parts in the contents of this thesis. The first part introduces a neural-fuzzy controller based on RBF neural network with application to the course control of ships. The second part describes a design method for a Data Interface System, which is a subsystem of the project named Integrated Bridge System (IBS) sponsored by the Chinese Ministry of Communication. The third part presents an energy-saving auto-explain&light control system used in museums.The first part of the thesis describes the theory of fuzzy-control and neural network-control, and also analyzes the characteristics of both the fuzzy-control and the neural network-control systems. The neural-fuzzy controller based on RBF neural network is introduced and applied to the course control of ships. The controller makes use of the function of non-linear mapping and the learning ability of RBF neural networks. The non-linear mapping is used by the controller to realize the fuzzy-control rules by training the RBF neural network offline. And then, the trained neural networks replace the fuzzy inference system in the online control system. An adaptive self-adjust unit is used to boost the system up and to reduce the static errors of the system. The thesis expatiates the design method and the realization of the controller and introduces the processes and functions used to train and to realize RBF neural networks. The method of constructing S-function to carry out the Simulink simulation is also given in the thesis. Simulation studies on the controller are conducted in many situations. Dynamic characteristics and robust stability of the controller are demonstrated by the simulation results. The controller has the potential of applications to many industrial systems.The second part introduces the design method of the Data Interface System, which is the subsystem of the project named Integrated Bridge System (IBS) sponsored by the Chinese Ministry of Communication. A intelligent CAN-PC bus card, a multi-port serial card and a 100/10M self-adjust network card are used to maintenance the stability and real-time capability of receiving and sending data in the communication subsystem. The technique of inquiring and interrupting is used to avoid the data-conflict in the software design.In the third part of the thesis, a system of energy-saving auto-explain&light control used in museum is presented. The system consists of an AT89C51 single-chip microprocessor and interface circuits. When visitors or audience come near, the sensorwill detects them and the system plays the commentary what is recorded in advance and in the same time, a special chip controls the lights from dark to light changing slowly. When no visitor there the lights become dark automatically, the energy-saving is achieved.