Transportation mode identification based on smartphone

Transportation mode surveys are essential resources in transportation research. The survey data is required for urban traffic planning. We want to be able to realize a way of traffic survey, in the case of without being limited by the condition, we can master everyone's transportation state in anytime and anywhere. In this paper, we develop a transportation mode survey application for the Android platform. We can obtain real-time information from GPS and accelerometer sensors and the real-time information will be stored in SD card with the form of database. We describe the features extracted from GPS and accelerometer sensors used to identify transportation modes with machine learning algorithms of decision tree. Experimental results show that the classification accuracy of the algorithm is 93.6782%. Finally, we select a test route and verify the reliability of the transportation mode identification system.