Download Full Text (390 KB)


The objective of this research is to provide reliable estimation of urban roadway travel time in time for traffic managing departments and travelers based on floating car data. Travel time data collection and estimation is an important technology method to achieve the Intelligent Transportation Systems (ITS) information services. Poor-quality information leads mistrust and un-ease of traffic congestion. Thus the accuracy of travel time prediction must meet certain requirements. GPS floating car collection method accesses the data source by ordinary vehicles equipped with positioning and wireless communication devices (such as taxis, buses, trucks, private cars, police cars, etc.), which provides more efficient, accurate and in-time data. There are two parts in my research: (1) The collection and pre-treatment for urban road travel time data with GPS floating cars, and (2) The estimation for travel time. I process and filter the data with the algorithm by clarifying abnormal fluctuations, losses, errors, and validation in the data set. As the similarities of the influence factors for travel time according to time periods and road sections, the mutation analysis is applied to divide the traffic flow data into traffic periods, such as rush hours, ordinary hours, and night hours. To predict short-term travel time, I employ BP neural network model based on local optimization. Travel time information reflects the state of underlying roadway traffic, predicts the duration of traffic congestion, and determines abnormal states. The estimated travel time in the next period on a certain road can be published to those who are in need. Also, pilots can choose the lower-traffic-flow roads regard the estimated travel time, which help shorten travel time and ease the congestion in rush hours.

Publication Date


Project Designation

Graduate Research

Primary Advisor

Ruihua Liu

Primary Advisor's Department



Stander Symposium poster

Road Travel Time Estimation with GPS Floating Car Data