Transportation Mode Recognition Algorithm Based on Bayesian Voting

• Published in: Proceedings of the ... Enterprise Systems Conference (Online) pp. 260 - 269
• Main Authors: Qin Yanjun, Jiang Mengling, Yuan Weichao, Chen Shaomeng, Luo Haiyong
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.09.2017
• Subjects: Acceleration; bias; classifier; Feature extraction; Global Positioning System; Gravity; low power consumption; Sensor phenomena and characterization; Transportation; transportation mode recognition
• ISSN: 2572-6609
• DOI: 10.1109/ES.2017.50

Nowadays, accurate identification of people's ways of transportation plays a major role in understanding users' mobility, analyzing and predicting traffic conditions as well as exploring new patterns of social activity. With regard to low power consumption, complex environment and other challenges, we propose a traffic model based on Bayesian detection algorithm in this paper. To cope with these challenges, we propose a Bayesian-based traffic pattern detection algorithm. This algorithm leverages variety of sensors embedded on smartphones to analyze and explores the different characteristics of various sensors to identify different traffic patterns. The results of massive experiments show that the traffic pattern recognition algorithm based on Bayesian algorithm has better universality and accuracy, with accuracy rate being over 91.5%.