Analysis of Link Travel Time Estimation Methods Based on Probe Vehicle Information and Road Segments

摘要

The link travel time estimation is currently a major issue since this value is used to find the fastest path from an origin to a destination. The probe vehicles system used to collect information about traffic is one of the most efficient methods in terms of cost and coverage. We propose new techniques to estimate link travel times using probe vehicle information and we also introduce the concept of road segments to play the role of low-pass filter. The road segments provide a means of improving the estimates by reducing the vehicles travel time spread along the link. In our simulations, we defined the accuracy of the estimates as being the difference between the link travel time estimated with the probe vehicles and the "true" link travel time. Furthermore, we considered the effect of varying the percentage of probe vehicles from 1% to 10%. The simulations show that the deviation for a probe vehicle ratio of 5% varies from 4% to 28%. Furthermore, we show that the number of vehicles and the traffic status are key factors in achieving accurate estimates.