Travel Mode Recognition Using Mobile Phone Signaling Data

Abstract

AbstractUrban residents can choose different travel modes, including walking, cycling, taking public transportation, and driving, and the information about the residents’ daily travel mode choices is important in the urban traffic planning and management. Therefore, the collection of such information is important. However, the traditional approaches are generally resource-consuming. Mobile phone signaling (MPS) data provides a chance to effectively collect residents’ travel mode choices. MPS data are generated when mobile phones connect to their nearby base stations. With the widely use of mobile phones, millions of MPS records are generated every day. This study explores the application of MPS data in travel mode recognition. Based on the MPS data and navigation trajectory, the proposed algorithm has four main steps: a) invalid data cleaning, b) staying point recognition, c) road matching process based on the Hidden Markov Model, and d) trajectory similarity calculation based on the Dynamic Time Warping. An Experiment is preformed to test the applicability of the proposed algorithm.KeywordsTravel mode recognitionMobile phone signaling dataRoad matchingNavigation trajectory