An Intelligent Controller Design for Automobile Anti-collision Based on Fuzzy Network

In order to solve the problem that the membership functions of traditional fuzzy control are difficult to be determined, a new method (fuzzy network) of fuzzy control is proposed. Fuzzy network adopted the network structure in which the membership functions, control rules, fuzzy reasoning and defuzzification process are included. Compared with the fuzzy neural network, fuzzy network don´t need learning and training, and the output control can be directly obtained after the data are inputted. Through the contrast of the design for automobile anti-collision system, the membership functions are not easy to be determined by the traditional fuzzy control method. Furthermore, the local surfaces are non-smooth. On the contrary, the membership functions are very easy to be determined by the fuzzy network method. In addition, this method can generate smooth surface. The simulation results of automobile anti-collision system indicate its validity.