Recognition Method of Motorized Traffic Mode Based on GPS Data
Konferenz: ISCTT 2022 - 7th International Conference on Information Science, Computer Technology and Transportation
27.05.2022 - 29.05.2022 in Xishuangbanna, China
Tagungsband: ISCTT 2022
Seiten: 4Sprache: EnglischTyp: PDF
Autoren:
Wang, Jizhao (School of Mechanical Engineering, Xinjiang University, Urumqi, China)
Guo, Dudu (College of Transportation Engineering, Xinjiang University, Urumqi, China)
Wu, Zhizhou (College of Transportation Engineering, Xinjiang University, Urumqi, China & College of Transportation Engineering, Tongji University, Shanghai, China)
Inhalt:
A clear understanding of the urban residents' travel behavior can serve as the foundation for traffic management and control measures. This paper proposes a mode recognition method for urban residents' travel behavior based on GPS data. With the GPS data preprocessing, feature fitting and machine learning, the algorithm is designed to recognize three common motorized traffic modes: bus, taxi and private car. The machine learning algorithm based on neural networks is used to recognize taxi and private car by learning the differences between peaking frequency, speed, acceleration and jerk. The average recognition accuracy is 83.84 percent, which shows this method has high recognition accuracy.