In order to improve traffic efficiency in urban road networks, the analysis and prediction of congestion is essential. In this paper a method is presented that first identifies frequently congested regions of a network by clustering affected links and then predicts travel time losses inside these clusters. Thereby, the reduction of the entire network to its mostly congested regions allows considering only a small set of highly relevant variables describing traffic congestion on a network level. For traffic prediction a K-Nearest Neighbors approach is developed that compares the severity of traffic congestion during morning peak with historical patterns in order to provide expected travel time losses during evening peak. Using the floating car data from a huge fleet of vehicles in Munich, Germany, the method is evaluated. Congestion clusters are identified and a study is conducted demonstrating the correlation between the severity of morning and evening traffic among different clusters. Finally, the performance of the predictor is assessed. Results show that the proposed approach achieves significantly better results than naive predictors.
Travel time prediction in partitioned road networks based on floating car data
01.11.2016
656871 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Travel time prediction method based on floating car data
Europäisches Patentamt | 2020
|Travel Time Prediction in Urban Road Networks
British Library Conference Proceedings | 1997
|Urban Road Travel Time Prediction Based on Taxi GPS Data
ASCE | 2013
|