Traffic Congestion Cause Identification Method for Urban Main Roads

Abstract

Identifying the causes of congestion is the key to solving traffic congestion. To improve the efficiency of congestion control, this paper establishes a congestion cause identification method based on the three stages of pattern recognition, source tracing, and cause discrimination. The K-means algorithm was proposed to calculate the frequency threshold of recurrent congestion, trace the sources of congestion according to the rules of congestion propagation time sequences, build a congestion fault tree based on causal logic relationships, and determine the occurrence probability and importance of each cause by using the expert scoring method and cloud model. The test results showed that the method is promising and could provide support for scientific congestion control.