01557065

Component

This paper proposes a model to express the propagation of backward forming and forward recovery shockwaves on a freeway. Unlike classic shockwave theory which is deterministic, the proposed model explains the propagation of shockwaves as a stochastic process. The state of the process is defined based on the traffic states of both downstream and upstream detector stations, and the probability of spillback or recovery is computed as a function of traffic measurements. Separate models for backward forming and forward recovery shockwaves are developed, and model parameters are estimated using the maximum likelihood method. The authors also propose a procedure to use the proposed model to improve the accuracy of near-future speeds predicted by time series models. The proposed procedure consists of two main modules: a) a time- series predictor which is used when the traffic condition is temporally constant, and b) a congestion detector. As soon as the congestion detection module detects that a station of the freeway is congested, the proposed stochastic shockwave models are activated to update the predictions provided by the time-series model. The authors apply the proposed procedure to 30 days of aggregated 5-minute loop detector data from a freeway in Toronto, Canada. The model is used to predict traffic conditions (speed) 15 minutes into the future. The results show that the proposed procedure improves the accuracy of the predictions by 17 -28% when traffic conditions are changing. The model is suitable for use in real-time freeway travel time or speed prediction applications.

This paper was sponsored by TRB committee AHB20 Freeway Operations. Alternate title: Stochastic Model for Predicting Shock Waves on Freeways.

01550057

15-2789

English

17p

2015

Transportation Research Board 94th Annual Meeting

Digital/other

Figures; References; Tables

Freeways; Mathematical prediction; Stochastic processes; Traffic congestion; Traffic flow; Traffic flow theory; Traffic queuing; Traffic speed

Toronto (Canada)

Highways; Operations and Traffic Management; Planning and Forecasting; I71: Traffic Theory

Transportation Research Board Annual Meeting 2015 Paper #15-2789

TRIS, TRB, ATRI

Dec 30 2014 12:57PM