Congestion Identification And Correlation Analysis On Urban Traffic Based On Taxi GPS Data

Accurate congestion identification and correlation analysis on urban traffic, which can sense road congestion and find the congestion rules, not only provide a reference for people to travel, but also provide a reference for related departments to take traffic guidance and traffic control measures. Existing traffic congestion research mainly identify the traffic congestion by getting a small scale of parameters such as velocity, density, traffic volume and vehicle delay using the fixed detection equipment,,which lacked the traffic recognition of the global road network, and can’t alarm the traffic congestion effectively,even avoid traffic congestion. So, the method to identify the traffic congestion on global road network is proposed in this paper, which realizes the rapid identification of traffic congestion. And the article analysis the spatial-temporal correlation of the urban traffic congestion.The main contents of this paper are as follows: First, identifying traffic regions in congestion needs scientific and effective indexes and criteria, so, the article uses mean velocity as the index of congested traffic regions on the basis of previous studies in this paper, and gives the process to calculate the index and the method to determine the index of the standard threshold on different traffic state through the fuzzy c-means clustering. Secondly, the article combines the urban road network with the characteristics of taxi travel, and proposes Apriori algorithm based on spatial-temporal constraint, which is used to analyze the correlation between the urban congested traffic regions. Finally, the article identifies the regions in traffic congestion by processing the GPS data in January 2014, extracting the speed of the taxi trajectory, calculating the average speed of the region, analysis the traffic congestion of working day and non-working day in Harbin, and finds the differences of traffic congestion states between working day and non-working day by the overlap rate of traffic congestion. Besides, the article mines the frequent item set of traffic congestion in January 2014 by processed spatial-temporal Apriori, extracts correlation of traffic congestion and analysis the spread rules of traffic congestion on working day and non-working day.The theory and the method of the identification on traffic congestion and the correlation analysis put forward in this paper serve to identify traffic congestion and prevent the traffic congestion, provide an effective method to identify and forecast the traffic congestion for the urban traffic management departments, contribute the extend of the theory on traffic congestion management, and has important practical significance to improve the management level of traffic congestion.