Urban Arterial Travel Time Estimation Using Buses as Probes

The accurate estimation of travel time of different types of vehicles in a traffic stream has always been of interest in various stages of planning, design, operations and evaluation of transportation systems. The traditional way of travel time data collection by means of active test vehicles or license plate matching techniques has its own limitations in terms of cost, manpower, geographic coverage, sample size and accuracy. With the growing need for real-time travel time data, the passive probe vehicles with onboard global positioning systems (GPS) are increasingly being used. However, due to privacy issues and participation requirements, the public transit vehicles are the only ones which can be equipped with GPS devices and this could possibly be used as a source to estimate the travel time of other types of vehicles. The present study is an attempt in this direction. Two approaches have been proposed: one based on the ratio of the section travel times of personal vehicles to public transit and the other based on the quantifiable relationship between the public transit and personal vehicles section travel times. The results showed that the approach-2 which is based on the relationship between the bus travel time and other vehicles travel time outperforms the approach-1, with 98% of the times the deviation of estimated travel time of personal vehicle with respect to observed/actual travel time being less than +/- 5 min and mean absolute percentage error (MAPE) within the acceptable range of 10-15%.