Traffic Detection Technology Based On Mobile Communications

Traffic control and guidance system is a core subsystem in ITS,and the key issue to implement this is how to make a real-time traffic flow detection and prediction.For the problem of NLOS in wireless location, the traditional TDOA /AOA algorithm is improved.The basic idea is the introduction of elastic factor to reduce the NLOS error, the simulation performance analysis shows that the improved algorithm does not need to know a priori information, effectively reducing the spread of non-line of sight impact,and the performance of location is improved effectively.Concerning the large data,requirement of real-time the paper investigates the dedicated Map-Matching algorithms.Considering the direction angle of data and the road information of electronic map, topology of road matching algorithm, and real-time positioning and accuracy is improved effectively.For the limitations of slow convergence of learning algorithms, We use L-M algorithm.A multi-BP neural networks model was constructed for the prediction of traffic flow parameters.LM algorithm has faster convergence speed than conventional BP algorithm.Compared with conventional forecasting methods, BP neural network meet the random parameters, in time to follow the changes of traffic parameters, so a higher accuracy, better adaptability. Simulation results show that the method is effective.