Passenger-to-Train Intelligent Matching Model Under Full-Length and Short-Turn Routings Based on Path Calculation from Map Application Programming Interface

Abstract

Aiming at flexible composition conditions, this paper proposes a passenger-to-train matching method that integrates automated fare collection (AFC) data and full-length and short-turn routing schedules. In the data layer, the method uses an electronic map application programming interface (API) to get path data in different modes to construct a workable path database. The traveling time of the AFC data is then used to match the path in the corresponding mode with the error of the corresponding database to get the passenger’s travel trajectory. In addition, the Passenger-to-Train Intelligent Matching System (T-P IMS) model is constructed based on the exploration of passenger transfer behavior in the microscopic stations of full-length and short-turn routing, and the projections of passengers taking trains are completed based on the train schedule. In this paper, the Behavior Analysis Intelligent Matching (BAIM) algorithm is also designed to realize efficient computation, and a real case study is carried out to count the average waiting time of transferring passengers as well as the train–passenger flow under different modes of operation. The statistical results of the case study show that the path matching effect of this paper is extremely close to the actual situation. The success rate of the model coupling is more than 99%.