Vehicle positioning and speed estimation based on cellular network signals for urban roads

Abstract

In recent years, cellular floating vehicle data (CFVD) has been a popular traffic information estimation technique to analyze cellular network data and to provide real-time traffic information with higher coverage and lower cost. Therefore, this study proposes vehicle positioning and speed estimation methods to capture CFVD and to track mobile stations (MS) for intelligent transportation systems (ITS). Three features of CFVD, which include the IDs, sequence, and cell dwell time of connected cells from the signals of MS communication, are extracted and analyzed. The feature of sequence can be used to judge urban road direction, and the feature of cell dwell time can be applied to discriminate proximal urban roads. The experiment results show the accuracy of the proposed vehicle positioning method, which is 100% better than other popular machine learning methods (e.g., naive Bayes classification, decision tree, support vector machine, and back-propagation neural network). Furthermore, the accuracy of the proposed method with all features (i.e., the IDs, sequence, and cell dwell time of connected cells) is 83.81% for speed estimation. Therefore, the proposed methods based on CFVD are suitable for detecting the status of urban road traffic.