[Objective]In order to accurately calculate the required scale of the connection facilities at Guangzhou urban rail transit stations,it is necessary to study and predict the sha-ring rate of passenger flow under each transportation connection mode at the stations.[Method]Based on the on-site investi-gations of the inbound passenger flow connection data of Nan-cun Wanbo Station,Tonghe Station and other stations in Guan-gzhou urban rail transit under different weather conditions,on the basis of the traditional MNL(Multinomial Logit)model,and in consideration of the impact of weather and differences in the inbound and outbound connection characteristics,an im-proved model for classifying the passenger flow transportation connection modes at urban rail transit stations based on the MNL model is constructed,and calibrated by using the data from the questionnaire surveys.[Result & Conclusion]The results of model calibration indicate that only the characteristic variable of connection distance passes the significance test,and there is no obvious correlation between factors such as gender,travel purpose and the choice of rail transit connection modes.The investigated survey data fails to capture the correlation be-tween the age of travelers and the choice of transportation con-nection modes.In the test of the improved model for classif-ying the transportation connection modes of the inbound pas-senger flow at Tonghe Station,on both sunny and rainy days,the passenger flow accuracy rates during the evening peak hours reach 86.0％and 77.2％respectively,showing that the im-proved model for the above scenario is superior to the tradition-al one.The improved model is applied to the target stations with similar land use attributes,confirming its effectiveness and rationality in actual passenger flow prediction.
urban rail transitpassenger flow at stationsclassification of transportation connection modesimproved model
万方数据
维普
