In order to improve the micro analysis and prediction of real-time forecasting method of dynamic parking demand, we selected three typical residential areas in Yangzhou City as an example to analyze the time-varying characteristics of motor vehicles’ arrival and departure. Considering the obvious difference between the arrival and departure characteristics of motor vehicle in residential areas on weekdays and weekends, the different time series models were used to forecast the berth occupancy of three residential areas on weekdays and weekends. Due to the higher proportion of commute travel on weekdays and the higher proportion of flexible travel on weekends, the variation tendency of berth occupancy on weekends is not as stable as that on weekdays. The result shows that the prediction accuracy of real-time numbers of berth on weekdays is usually higher than that on weekends. On weekdays, the berth occupancy rate of three residential areas is regular, which can be forecasted by ARIMA (Autoregressive Integrated Moving Average) model, and can reach more than 98% of the prediction accuracy. Oppositely, the weekends’ time-varying regularity of berth occupancy is not obvious, thus using ARMA (Autoregressive Moving Average) model, and the accuracy can reach over 95%. Overall, time series model has good adaptability to the residential area, and the higher accuracy can be achieved by selecting the appropriate model.
Time-Varying Characteristics and Forecasting Model of Parking Berth Demand in Urban Residential Areas
Lect. Notes Electrical Eng.
2020-03-24
14 pages
Article/Chapter (Book)
Electronic Resource
English
Dynamic parking berth demand prediction method based on berth sharing
European Patent Office | 2022
|Dynamic demand-based berth parking mode management method
European Patent Office | 2024
|