A method for urban road traffic congestion identification based on GPS data

The effective recognition of urban road traffic status is of great significance to urban road intelligent traffic management. To solve the problem that a single indicator of track data cannot accurately identify road traffic congestion, this paper defines three-dimensional traffic flow indicators of trajectory speed, trajectory traffic volume, and trajectory density according to the sampling characteristics of GPS data, and proposes a road traffic congestion status recognition method based on vehicle track data. Firstly, based on the road network modeling, road segment matching is carried out, and three traffic flow indicators of each road segment were calculated with 5 min as the time granularity. Then, the K-means algorithm is used for cluster analysis to obtain four state categories: unblocked, basically unblocked, general congestion, and severe congestion. Secondly, using the Genetic Algorithm and Mixed Parameters to optimize the Multi-classification Support Vector Machine, the GA-MP-MSVM model for road traffic state recognition is proposed. Finally, an experimental analysis is carried out based on the urban road network of Shenzhen and the GPS trajectory data of taxis. The results show that the constructed traffic flow indicators can effectively distinguish the traffic congestion status of the road, and have a good recognition effect on the constructed GA-MP-MSVM model, with a recognition accuracy of up to 99.53%, which is 3.74% higher than that before optimization.