Fuzzy control of automatic automobile obstacle avoiding

Abstract

The unmanned automobile has become a major research project in the automobile industry. The automobile will meet obstacles unavoidably on the road, so it is necessary to design a feasible controller for avoiding obstacles. Because of the complexity of the automatic obstacle avoiding system itself and of the uncertainty of the parameters, the original method of designing a controller by establishing models is limited. Driver experience provides a good control example for designing controllers and one can make use of this experience to design a controller for avoiding obstacles, which simulates the driving behaviors of humans. At present, fuzzy control technology has been widely used in many fields. It is commonly considered as an effective method to process uncertain information and control complicated nonlinear systems. It does not rely on the model of the controller object, with a better adaptability to changes in parameters of the system. The expression ability of the fuzzy language variable can be used to describe the experiences of humans. However in actual applications, where there are more fuzzy variables, the rule base will have a very large scale and the regulation of rules will become very complicated. In this article, a fuzzy controller based on a neural network is put forward through analyzing the features of the fuzzy control model and the driving experience of humans, greatly simplifying the design of the fuzzy controller for complicated systems.