Investigation Of Vehicle Geomagnetic Detecting System And Recognizing Algorithm

At present, the automatic recognition of automobile type is the hot topic to support Intelligent Transportation System(ITS) and a core technology in the urban intersection traffic lights optimize control system and the highway charge system. Obviously, this research work has practical meaning in future ITS engineering. According to the demand from key project by Wuhan Science and Technology Institute, a intensive explore and research on recognition methods is presented based on current vehicle recognition development status and the traffic optimize control demand, we break through the conventional fuzzy method and adopt multi-data fusion method based on fuzzy logic considering the features of vehicle geomagnetic signals, a vehicle classification system is designed in this paper using fuzzy data fusion method which fully representing some specific advantages like low cost, high credibility by using AMR sensors in the system.In this paper, an effective vehicle classification method is presented. The features are firstly picked up according to the ambipolar property of the vehicle geomagnetic curves and then the vehicle type model functions are built according to the sample vehicles, at last, the detected vehicle could be recognized using multi-data fusion methods. This method could be also applied in large parks so as it has a great application foreground.The principal part of this paper shows the theory and the applications of AMR sensors in traffic detection, and the study status of vehicle classification using fuzzy logic is detailed, further, this paper introduces a new Fuzzy Data Fusion arithmetic in vehicle identification and implements it in practical system.In the end, a traffic detection experiment is fulfilled by the identification system and the veracity of five different data fusion methods are compared by analyzing the sample data. Finally there is a conclusion of their advantages and disadvantages, as well as the improving methods and the developing direction in future.