A Travel Mode Identification Framework Based on Cellular Signaling Data

The rapid development of telecommunication network has produced a large amount spatial-temporal information mobile phone users. GPS data are typically collected by smartphones apps, which restricted to small samples the population. Cellular signaling (CSD) usually operators, enables researchers conduct travel behavior analysis entire population at relatively low cost compared GPS. However, extracting mode from CSD is particularly challenging due noise and positioning accuracy. This paper proposes identification framework based on CSD, includes cleansing identification. In terms cleansing, oscillation sequence drift mainly cleansed. For sequence, this detection algorithm time window. data, distance, velocity frequency. identification, task divided into two dichotomous problems: motor non-motor transport public private Each problem an that does not rely ground truth dataset for model training. Finally, constructed verify all algorithms. result shows that, in similarity between after actual trajectory according DTW improved 101.13% average. proposed method can achieve similar or even better accuracy than traditional supervised-learning algorithms (94% 83.5% identification), be directly applied large-scale scenarios.