An algorithm for prediction of travel time using vehicle sensor data

Travel time is usually calculated by dividing a route into several subsections and totaling travel time measured by vehicle sensors in each subsection. The disadvantage of this method is that the vehicle length is always assumed to be constant despite changes in traffic conditions. The paper proposes a new approach which introduces a time series model and a statistical presumption of vehicle length. Using the former, one can predict changes in traffic conditions from origin to destination during the passage of time. With the latter, one can calculate travel time from the point of origin. To evaluate the validity of the approach, the author conducted an experiment on national road Route 16, in Chiba Prefecture, Japan.<>