Empirical study of dynamics pattern of passenger transit trip based on time characteristics

The modern transportation system consists of people, vehicles, roads and information concerning them. For a long time, the study of urban public transportation mostly focuses on the mathematical modeling of passenger flow characteristic while ignoring the internal mechanism from a disaggregated level, especially the subjective initiative of passengers, the behavior subjects. The development of information technology and application of technologies such as complex networks and behavioral dynamics provide data and technical support for analyzing passenger transit trip patterns and studying the driving mechanism of transit trip behaviors. The time interval distribution for passenger transit trip was verified to be power-law distribution at the group level and individual level, respectively, and the power-law distribution was also observed to disappear in some individual behaviors. The inherent burst trait of human behaviors made the power-law universal, and breaking the universality precisely revealed the deeper hidden laws behind, namely passenger transit trip pattern. In addition, empirical study at the individual level showed that behavioral dynamic parameters about passengers’ transit time interval distribution such as the power-law fitting coefficient, burst and memory cannot be used for passenger transit trip pattern recognition. Based on the analysis of demand driving trip, a recognition method of passenger transit trip pattern based on image recognition is discussed in this paper.