Modeling of taxi drivers’ experience for routing applications

Taxi drivers drive on the roads every day and become very knowledgeable of the spatiotemporal traffic patterns in a city.It therefore is reasonable to assume that the routes chosen by taxi drivers often work out better than those selected by other drivers.Since dynamic navigation assistance based on real-time traffic information faces limitations such as the spatial coverage of real-time data collection sites,performance of real-time data processing and communications,and accuracy of short-term traffic forecasts in a large urban area,experiences gained by taxi drivers can be a valuable data source for improving the quality of vehicle navigation guidance.This paper develops a vehicle navigation guidance system based on taxis drivers’ knowledge derived from floating car data collected over an extended time period.We then classify road segments based on the spatiotemporal characteristics of taxi tracking data.A case study using taxi tracking data collected in Wuhan,China is presented in this paper to demonstrate the performance of this vehicle navigation system based on taxi tracking data.