Based Access To Freeway Traffic Flow Parameters Of The Ultrasound Study

As the traffic issue stands out as a global problem, Intelligent Transportation System (ITS) becomes a heated topic in current studies. ITS has a wide extension, and includes a lot of subsystems. This paper is focused on the field of the traffic parameters acquirement which lies on the bottom of the ITS. After comparing and analyzing lots of conventional traffic detecting schemes, this paper designs a detecting scheme based on ultrasonic. This scheme is based on the distance data obtained by ultrasonic telemeter.This paper does a lot of works on the research of the acquirement of traffic parameters, and has realized a series of algorithms, which include a kind of vehicle segment algorithm which is insensitive to the noise, a kind of wave matching algorithm based on the time sequence and the algorithms to compute the vehicle height, vehicle length, vehicle speed, traffic load and density. According to the features of the ultrasonic waves, the vehicle segment algorithm programmed in this paper can extract the vehicle waves exactly under the disturbance of the system noise, and the wave matching algorithm programmed in this paper is robust against the system noise which comes from the vehicle segment algorithm. When matching error occurs, the enor will not be accumulated to influence the matching process of the next vehicle.In addition, this paper emphasizes the research of vehicle type classification, and presents a universal vehicle type classification scheme, and introduces Bayes decision-making theory and Fourier descriptor model into the classification algorithm. The classification algorithm implemented in this paper is a multi-level classification algorithm, which uses Bayes decision-making theory to classify the vehicle into some main classes at first, then uses Fourier descriptor model to sort in deeper level. This paper uses Parzen-window method to acquire the condition probability density, and does particular research on the relationship of the sampling number and the describing ability of Fourier descriptor.The traffic parameters acquirement prototype system has already implemented on PC, and achieves ideal result in experiment.