Study On Short-term Sectional Passenger Flow Forecasting Method Of Urban Rail Transit Based On State Space Model

Urban rail transit (URT) characterized by large carrying capacity, high speed and comfortable feeling has become an effective means to relieve traffic congestion in major cities. Along with the constant improvement of urban rail transit network, great changes of the time and space distribution characteristics about passenger flow have taken place. Therefore, forecasting short-term passenger volume of rail transit lines reasonably and accurately is the very vital basis for operations management. The actual variation of urban rail transit and needs of passenger is reflected by sectional passenger flow, that is an evaluation of URT service levels and importance for emergency response. Operation and management departments of URT to develop and adjust operational plans are dependent on accurately predict of short-term sectional passenger flow.The current research situation of traffic forecasting at home and abroad are summarized, especially sectional short-term traffic forecasts. Most current projections based on a single target as the research object, can’t reflect the interplay between sections in time and space. In view of this situation, short-term sectional passenger flow forecasting model for URT is presented based on the correlation analysis between sections.Firstly, the passenger flow development trend of URT is analyzed, and the volume development trends of Beijing railway transit is detailed analyzed. Sectional passenger flow distribution characteristics of different types of urban rail transit line are respectively studied from two aspects of time and space.Secondly, multidimensional scaling method is introduced to carry out correlation analysis between the urban rail transit sections. Sections of urban rail transit lines are divided into several related groups, according to the correlation of time and space. The related groups are analyzed based on evolution mechanism of the sectional passenger flow. The correlation of sections is verified. That laid foundation for building a short-term passenger flow forecasting model, considering time-space change of urban rail transit sections.Then, on the basis of studying the correlation of urban rail transit sections, state space model is combined with time series analysis to set up a short-term passenger flow forecasting model of multi-sections based on state space model. Multi-sectional short-term traffic forecasting model and method based on state space model is proposed built. The state space model with time time-varying coefficient and the state space model section passenger flow as state variables are built, two-stage adaptive Kalman filtering method is used to solve the model and at the same time considering the influence of noise covariance on forecast accuracy, the adaptive estimation of noise covariance is conducted by time-varying noise estimator. And the measure is taken to control the divergence of filter.Finally, according to the actual section passenger flow data of a rail transit line, the case analysis is used to test the effectiveness of the multi-section short-term passenger flow forecasting method. The results show that, the prediction accuracy of the forecasting model considering the correlation between sections is superior to the model just consider a single section, and get all predicted values of the relevant sections of the group. The proposed forecasting method has better applicability and rationality.