Traffic State Determination System Based On RealTime Traffic Parameters Detection

Traffic state is mainly represented by the traffic flow state, which mainly reflect to the level of the road traffic congestion. Determining of the road traffic flow state in real time and accurately is the precondition and basis for traffic control system. Former research on traffic state determination is always about fast road and highway. As the problem of urban traffic is more serious than before, urban road traffic state determination becomes a hotspot in research area of ITS. A wireless sensor network model for traffic flow detection is present in this thesis Which is mainly about how to detect traffic parameters precisely and improve the method for traffic state determination.First, the shortcoming of the existing technology in traffic flow parameters detection is demonstrated. Magnetic sensor can effectively compensate for these shortcomings. The detection technology of traffic flow parameters based on WSN is involved. The research was first raised in the PATH report of University of California, Berkeley. The report proposed a 2 node model which is used in vehicle speed detection. So 3-nodes traffic flow parameters detecting model has been proposed. The results of the experiments are demonstrated that the vehicle speed captured by the 3 nodes model is more precise than 2 nodes detection model. In order to meet the characteristics of long-term deployment in traffic flow parameters detection, we presented a power consumed scheme Duty-cycling-V. The scheme can dynamically adjust the working cycle of the sensor node according to the speed of traffic flow. The results of the experiments demonstrated that the scheme has better power efficiency.Second, urban transportation is distributed, man-involoved and time-variable system, has spatial-temporal and stochastic characteristics. In this thesis, the FCM has been used to extract the traffic state characteristics, and then the characteristics and real-time traffic parameter are used to determine traffic state. As the FCM algorithm is an unsupervised clustering algorithm. The useful information in its implementation process is ignored. The new reducing dimension data can be used to go on clustering and extract data feature rapidly. Its validity and correctness is validated by the experiment.