2020/09/17 by Felix Rempe, Klaus Bogenberger, Rempe, Felix +1
Engineering · Social Sciences · #Traffic Prediction and Management Techniques #Transportation Planning and Optimization #Vehicle emissions and performance
paper · pdf · doi:10.48550/arxiv.2009.08354
Most traffic state forecast algorithms when applied to urban road networks\nconsider only the links in close proximity to the target location. However, for\nlonger-term forecasts also the traffic state of more distant links or regions\nof the network are expected to provide valuable information for a data-driven\nalgorithm. This paper studies these expectations of using a network clustering\nalgorithm and one year of Floating Car (FCD) collected by a large fleet of\nvehicles. First, a clustering algorithm is applied to the data in order to\nextract congestion-prone regions in the Munich city network. The level of\ncongestion inside these clusters is analyzed with the help of statistical\ntools. Clear spatio-temporal congestion patterns and correlations between the\nclustered regions are identified. These correlations are integrated into a K-\nNearest Neighbors (KNN) travel time prediction algorithm. In a comparison with\nother approaches, this method achieves the best results. The statistical\nresults and the performance of the KNN predictor indicate that the\nconsideration of the network-wide traffic is a valuable feature for predictors\nand a promising way to develop more accurate algorithms in the future.\n