Research On Quickly Recognition Ways Of The Urban Road Non-Recurrent Local Congestion

Because of the traffic have huge effects, it occupies a position cannot be ignored, but the construction of urban transportation infrastructure is building more slow, it leads to the urban traffic already cannot satisfy the needs of people strongly. Traffic congestion has become a problem urgently to be solved in the city. Recurrent congestion have a regular, we can recognize them by the law of the daily statistics. And incidental congestion is generated randomly, it has the quite big chance, it is difficult to determine the occurrence of time and place, it bring very great difficulty for traffic evacuation, the traffic incident happened caused great threat on people’s life and property security caused great threat. Many domestic and abroad scholars carried out extensive research for the identification and control of the accidental congestion, but the effect is not beautiful, because of identify was limited by the quickness and the effectiveness. Therefore, it is necessary for starting the recognition of accidental congestion problem, in order to control congestion in a timely manner.In this paper, on the basis of many previous studies of traffic, We conduct the research on accidental congestion. First I has carried an overview of the urban road traffic congestion, described the meaning of traffic congestion and analyzed the causes of traffic congestion in a certain extent, analyzed the mechanism of urban traffic congestion, introduced several kinds of traffic flow parameters, comprehensive analyzed the characteristics of traffic flow and traffic network features. To focus on several classic algorithm for recognizing traffic congestion and judge conditions for algorithm, comprehensive analyzed and compared recognition performance of those algorithm. Finally put forward the concept of BP neural network on the basis of artificial neural network, through to the improvement of network model for recognizing urban road accidental congestion, using the data collected to start traffic to visit the true through traffic simulation VISSIM software, we can get the traffic flow, average speed and share, its input to the network model, to be used for the recognition of accidental traffic jams, through the example analysis shows that the method has high accuracy, and has strong practicability.This article through to the in-depth study of urban road accidental local congestion, It would be helpful to accidental congestion control management and formulate reasonable evacuation methods.