Study On Urban Traffic Congestion Status Identification Based On RFID Data Of Vehicles

Traffic congestion refers to the fact that the traffic demand cannot be satisfied because the number of motor vehicles in the road approaches or exceeds the maximum capacity of the road.As a result,the phenomenon of stagnation of motor vehicles on the road occurs,and traffic congestion will affect people’s travel efficiency.This will bring a series of problems such as environmental pollution,traffic accidents.The identification of traffic congestion status is the basis for realizing traffic control and traffic guidance.On the one hand,traffic congestion status can make traffic managers understand the running state of road network traffic,also,it provides the prerequisite for traffic information service functions such as route planning and recommendation.Generally speaking,urban road traffic congestion status identification can achieve effective traffic control on urban roads,reasonable traffic guidance to road traffic and reasonable planning of traffic network,which greatly improves the efficiency of residents’ travel and ensures the safety of residents’ travel.At present,urban road traffic congestion identification is mainly based on automatic discriminant technology.With the continuous development and improvement of the technical theory of traffic field,technologies such as fuzzy theory,neural network and interdisciplinary theory have been increasingly used in the study of traffic congestion discrimination.In addition,with the continuous advancement of traffic flow data collection technology,the use of more objective and accurate acquisition technology brings the research of traffic congestion identification a growing room for development.Radio Frequency Identification(RFID)has developed rapidly in recent years.This technology has been used in the field of traffic to realize vehicle monitoring,vehicle counting,vehicle discrimination,traffic congestion status detection and other important functions.The advantages of RFID are reflected in the accuracy of recognition not affected by weather and other factors.Also,vehicle recognition speed is faster and vehicle identification information is more comprehensive.This paper takes the RFID data of vehicles in Chongqing as the research object.A set of complete theory to analysis and determinate the congestion status of road traffic in Chongqing and it provides decision support for traffic management and travel planning.The main research content of this article is as follows:(1)Selecting Reasonable traffic flow parameters through RFID data of vehicles.These parameters including passenger car unit,traffic,speed,density.And select suitable parameters for evaluating traffic congestion status and input dimension of FCM fuzzy clustering algorithm.(2)Determine the optimal number of clusters for the FCM algorithm.The difference of traffic congestion in different roads corresponds to the difference of cluster centers of different sample data.Based on this difference.Based on this difference,we need to study the optimal clustering number of each sample data in order to ensure the optimal clustering effect and meet the actual traffic congestion state of road traffic.(3)Complete the mapping relationship between cluster centers and traffic congestion.For each clustering result,we can not know which type of traffic congestion in each category,so we need to establish a mapping model to associate each cluster center with specific traffic congestion status.In this paper,the cluster center traffic congestion status recognition model is established according to the rules of "ordering center of projection points in clustering" and the principle of "continuous left-most state matching".(4)Due to the shortage of the accuracy and efficiency of the traditional FCM algorithm,this paper use the generalized balanced FCM algorithm and the incremental FCM algorithm to optimize.(5)Completion of incremental FCM on the Storm platform algorithm parallelization.Combining the incremental FCM clustering process with Storm to achieve parallelization of the algorithm.