Short-term OD estimation for urban rail transit based on spatio-temporal characteristics
Accurate short-term origin-destination(OD)estimation technology is the main basis for urban rail transit operation organization plan development,and also the basis for urban rail section passenger flow prediction.To simultaneously consider the spatio-temporal dual dimensional features and achieve high-precision estimation of the entire network OD quickly,this paper analyzes the OD distribution characteristics of urban rail passenger flow based on both temporal and spatial dimensions.Then,the optimal input factor screening method under multiple dimensions is proposed.Furtherly,a multi-level nested OD estimation method based on multiple linear regression models is constructed.Numerical cases based on real-world data from Nanjing metro line network in 2019 are conducted.The results show that the mean absolute percentage error(MAPE)of the whole network OD are about 37%and 24%under the granularity of 15 min and 60 min,respectively.MAPE of some OD pairs is less than 5%during the peak hours.The root mean square error(RMSE)for a granularity of 15 min is approximately 1.3.Compared to existing short-term OD estimation methods,the model proposed in this paper has better prediction effect in peak hours,which is suitable for network-wide and time-wide OD estimation.
urban rail transitorigin-destination flowshort-term passenger flow estimationmultiple linear regressionspatial-temporal characteristics
国家科技期刊平台
NSTL
万方数据
