Integrated Prediction Model for Urban Rail Transit Station Feeder Passenger Flow

Abstract

[Objective] The rational coordination of feeder facilities is crucial for maximizing the functionality of urban rail transit. To quantitatively describe the distribution of passenger flows originating from urban rail transit stations and the mode share of feeder transportation, an integrated prediction model for feeder passenger flow is proposed. [Method] The current research status on domestic and international studies related to source region delineation, station selection models, and feeder mode selection is reviewed. The above integrated prediction model is established, consisting of feeder mode selection and station selection models, and both employ a nested Logit model. The Logsum from the ROOT layer of the feeder mode selection model serves as a connector between the two models. [Result & Conclusion] The Logsum output from the ROOT layer of the feeder mode selection model is utilized to characterize the comprehensive competitiveness of various feeder modes for spatial points, offering a more comprehensive reflection of overall costs. The station selection model is calibrated using discretized mobile signaling data, and in the case study of metro stations in the central area of Nanjing, the corrected pseudo-R2 for the calibrated station selection model reached 0.872, indicating a satisfactory explanatory performance. Comparison between the research case results and survey data reveals that the proposed integrated prediction model effectively describes the distribution of feeder passenger flows and quantifies the usage of different feeder modes.

Keywords

• urban rail transit
• station
• feeder passenger flow