Study On The Method Of Real-time Identification And Propagation Estimation Of Traffic Congestion On Urban Road

With more and more serious traffic congestion, urban road traffic congestion becoming a major cities plagued social problem, which seriously affect the city's sustainable development and people's daily work and life. It was from the study of theory and model of congestion, rapid and accurate identification of traffic state and timely information released, that the driver can be noticed with the best path to avoid traffic congestion; it also can ease the traffic flow, to avoid traffic gridlock caused. Traffic propagation theory and model play an important role on mitigation measures for congestion and have important theoretical value and practical significance.The study of discrimination on the congestion status, propagation rules and its model more and more thorough at home and abroad, but to achieve real-time traffic congestion state identification and timely and effective traffic incident optimal rescue operation, the state must first establish an effective discriminated model to get effective parameters of traffic flow. Considering evaluation of traffic congestion and fuzziness of the traffic state and subjective characteristics, this paper adopt clustering method for processing the traffic flow parameters, then build an adaptive neuro-fuzzy inference system (ANFIS) which can identify traffic congestion state. This paper analyzes the propagation rules and divided traffic congestion into individual sections and regional road network. Because the causes of traffic congestion in different, it's into recurrent congestion and non-recurrent congestion, and take a different method to estimate congestion propagation, the queue length and congestion during time.To verify the feasibility of identification system and the methods of congestion propagation estimate, this paper use of Paramics, a microscopic traffic simulation software for simulation experiment. After collection the traffic flow data and analysis to input the system. Then, the experiment results show that the improved method has high practicability.and feasibility, more accuratly. Compared with other classical algorithm, the improved has lower false identification rate and more identification accuracy. Simulation on traffic congestion propagation, respectively, for individual sections and regional road network, the results showed that the main road on recurrent congestion have heavy traffic load and propagation wide. Due to the main road have more lanes, more intersections and more streaming capacity, non-recurrent congestion showed dissipated fast, short duration and propagation narrow.