科研之友
微信
新浪微博
Facebook
分享链接
摘要
Estimation and prediction of urban travel time are acknowledged as important yet challenging topics. The traffic flow theory as developed especially for freeway traffic does not give much help for modeling urban traffic processes. Urban travel times are irregular because of several disturbances on a path. The interruption of trips at signalized intersections causes a large part of the total delay that vehicles experience on a route. These delays vary with stochastic properties of traffic flow, including stochastic arrivals and departures at signalized intersections. As a result, a wide distribution of delay (travel times) can be found for a certain traffic condition and traffic control scheme. This paper proposes a procedure for predicting the distributions of urban link travel times. The core of this procedure is the proposed model of distributions of link travel times that considers stochastic arrivals and departures at intersections and the traffic control scheme explicitly. A comparison of the distributions of link travel time predicted by the model with those derived from VISSIM simulation showed that travel time distributions can be predicted well under time-varying demand and different traffic conditions. A comparison with real-life data derived from Global Positioning System information indicates that distributions of link travel times predicted by the model can represent these real-life travel time distributions, though some discrepancies can be observed for certain links.
