Transportation mode recognition algorithm based on mobile phone signaling HMM trajectory mapping
CSTR:
Author:
Affiliation:

Key Laboratory of Transport Industry of Big Data Application Technologies for comprehensive Transport (Beijing Jiaotong University), Beijing 100044, China

Clc Number:

U491.1

Fund Project:

  • Article
  • |
  • Figures
  • |
  • Metrics
  • |
  • Reference
  • |
  • Related
  • |
  • Cited by
  • |
  • Materials
  • |
  • Comments
    Abstract:

    To address the limitations of sparse mobile phone signaling data and the insufficient ability to extract and fuse features in trajectory reconstruction, this paper proposes a precise transportation mode recognition algorithm based on trajectory reconstruction and multi-source feature fusion. A wireless signal loss model is developed to characterize the signal path loss and signal strength of base stations. Using the hidden markov model (HMM), the mobile phone signaling trajectory is reconstructed from a base station sequence to a road segment node sequence. A trajectory reconstruction method based on the wireless signal propagation hidden markov model (WP-HMM) is proposed to describe the relationship between signal strength and distance. Based on the reconstructed travel trajectory and combined with the characteristics of road segment types, a spatiotemporal standardized similarity measurement fusion navigation trajectory feature is proposed, and a transportation mode recognition algorithm based on random forest (RF) is constructed. Empirical analysis shows that, through the reconstruction of travel trajectories, the model′s average recognition accuracy improved by over 8%, and it demonstrated excellent generalization ability for new samples. Compared to existing methods, the spatiotemporal normalized similarity measure more accurately captures the movement patterns between trajectories. In trajectory recognition across different environments, the model performs significantly better in suburban areas than in urban areas. The proposed algorithm demonstrates significant application value in the field of transportation mode recognition using large-scale mobile phone signaling data.

    Reference
    Related
    Cited by
Get Citation
Related Videos

Article Metrics
  • Abstract:
  • PDF:
  • HTML:
  • Cited by:
History
  • Received:November 08,2023
  • Revised:
  • Adopted:
  • Online: October 19,2026
  • Published:
Article QR Code