Heuristic Analysis of Transport System Efficiency Based on Movement of Mobile Network Users

Abstract

The paper describes results of introductory research focused on possibility to use location data available in a mobile network for the analysis of transport system status and efficiency. The details of a system capable of detecting abnormal traffic situation (accidents, heavy congestion) are described. This system (called VASTAR) uses a neural network to learn and store certain characteristic of the analyzed part of a road system. Based on a measured divergence from normal characteristic, a notification about non-typical situation is triggered. The results of a computational experiment using real-world location data and simulation of abnormal situation are provided. The proposed system can be a relatively low cost way to improve competitiveness of a mobile network operator by allowing him to offer new type of informational service. It could also aid municipal authorities by providing support for decisions regarding road traffic control and management and be used by emergency services as a monitoring an alarming tool for detecting abnormal road traffic situations when other means of observation are unavailable.