On Estimation of Link Travel Time in Floating Car Systems

Abstract

Estimation of link travel times from measured values is a critical service of floating car information systems (probe car information systems), and the results of estimation are utilized to improve quality of vehicle navigation. Because actual link travel times are affected by various factors, such as time period, traffic volume, and weather, these factors should be considered to estimate the link travel time. Also, they should be considered to estimate the minimum amount of data needed to estimate link travel time while maintaining accuracy to optimize the floating car systems. In the systems currently being implemented, however, effects of the weather are not utilized to estimate the link travel time, and the minimum amount of necessary data has not been studied in detail. This paper reveals the effect of weather differences on estimated link travel time using data actually obtained by floating car systems. In particular, this paper pays attention to effects of snowfall. This paper also evaluates the amount of data required to obtain estimated link travel times considering effects of various factors.