Uncertainty and Predictability of Urban Link Travel Time

Abstract

136 ply (e.g., due to incidents, road work, weather conditions, and road geometry) on freeways; on urban arterials, travel time is significantly influenced by-besides fluctuations in traffic demand and supplytraffic control at, stochastic arrivals to, and departures from the intersection. Many of these factors are stochastic, which results in variable travel times. However, a complete and valid model to predict travel times that takes into account all these stochastic influencing factors seems unfeasible so far. In addition, most research done toward examining the variability in travel times has concentrated on long-term variability (e.g., peak hours, off-peak hours, and daily and seasonal variability). Traffic control has a special effect on the predictability of travel times. Two consecutive vehicles that enter the network at nearly the same time can have completely different travel times when the first vehicle just passes an intersection at the end of the green phase and the following vehicle has to stop. This may have an impact on delays at the next intersections, so that the first vehicle may have a much shorter travel time than the second one. This phenomenon, where small differences in the initial state can have large differences in the final condition, is called "bifurcation." Systems with bifurcations are ill-predictable.He et al. (7) examined the short-term temporal-spatial variations and correlations of travel times among urban links by using vehicletracking data. The results indicated strong evidence of significant correlation between link travel time and arrival time. The variability of travel time as a function of traffic demand was proposed by Liu et al. (8). Individual travel times were used to construct travel time distributions at different traffic demand levels. The fuzzy k-means method was applied to classify traffic patterns based on traffic flows. One limitation of this method is that it is data driven, and there is no analytical analysis to support the underlining traffic process on the urban road network. As Viti and van Zuylen (9) discussed, the stochastic delays at the signalized intersection constitute a large part of travel times on urban links. Understanding of the vehicle delay evolution or delay variability at signalized intersections can lead to more insights into the variability of urban link travel times and give more possibilities for travel time estimation and prediction.Delays at signalized intersections have been widely investigated, and researchers have taken great effort to approximate the analytical expression for the mean delay [Webster and Cobbe (10), Miller (11), Newell (12), and Akcelik (13)]. Much less work has been done to quantify delay variability at signalized intersections. Investigation into the delay variability can be done in two main ways: (a) by modeling the variance of delay and (b) by obtaining the delay distribution. To quantify the delay variability in an analytical way, Fu developed a model to predict the variance of overall delay (14). The methodology...

On the estimation of arterial route travel time distribution with Markov chains

“…They assume a binomial distribution for number of probe vehicles and a Poisson distribution for reporting rate, deriving formulas for mean and variance of reports number and speed estimation, and confident reporting intervals. Delay at signalized intersections is the main source of uncertainty in urban TTD which is tackled in (Zheng and Van Zuylen, 2010). The authors propose an analytical method for estimation of an urban link delay distribution.…”

“…two or three times the cycle length, can significantly be influenced by the size of study period. For example, choosing the start or end of the study period at the beginning or end of the cycle could cause drastic changes in the results (Zheng and Van Zuylen, 2010). However, by analyzing longer study periods; we can smooth out the traffic variations caused by time-dependent signal capacity.…”

Reconstruction of delay distribution at signalized intersections based on traffic measurements

“…From this point of view, using the travel time distribution instead of the mean travel time is a more realistic approach for travel time prediction in the urban road network. An analytical delay distribution model has been developed by authors [20]. The applications of the delay (travel time) distribution are two folds.…”

Analysis of delay variability at isolated signalized intersections

“…Variability can result from the differences in the mix of vehicle types on the network for the same flow rates, differences in driver reactions under driving conditions (Li and Shi, 2006;Jin et al, 2011), differences in delays experiences by different vehicles at intersections, and such random incidents as vehicle breakdowns and signal failures, etc. On urban arterials, delays incurred at signalized intersections account for a large part of travel time (Zheng and van Zuylen, 2010). The interpretation of delay evolvement or delay variability at intersections will help give a more comprehensive insight into arterial travel time variability, and provide more possibilities for travel time estimation.…”