Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2603.00340 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908895603064832 |
|---|---|
| author | Zhang, Yuandong Echchabi, Othmane Feng, Tianshu Zhang, Wenyi Liao, Hsuai-Kai Chang, Charles |
| author_facet | Zhang, Yuandong Echchabi, Othmane Feng, Tianshu Zhang, Wenyi Liao, Hsuai-Kai Chang, Charles |
| contents | Transportation mode detection is an important topic within GeoAI and transportation research. In this study, we introduce SpeedTransformer, a novel Transformer-based model that relies solely on speed inputs to infer transportation modes from dense smartphone GPS trajectories. In benchmark experiments, SpeedTransformer outperformed traditional deep learning models, such as the Long Short-Term Memory (LSTM) network. Moreover, the model demonstrated strong flexibility in transfer learning, achieving high accuracy across geographical regions after fine-tuning with small datasets. Finally, we deployed the model in a real-world experiment, where it consistently outperformed baseline models under complex built environments and high data uncertainty. These findings suggest that Transformer architectures, when combined with dense GPS trajectories, hold substantial potential for advancing transportation mode detection and broader mobility-related research. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_00340 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Detecting Transportation Mode Using Dense Smartphone GPS Trajectories and Transformer Models Zhang, Yuandong Echchabi, Othmane Feng, Tianshu Zhang, Wenyi Liao, Hsuai-Kai Chang, Charles Machine Learning Transportation mode detection is an important topic within GeoAI and transportation research. In this study, we introduce SpeedTransformer, a novel Transformer-based model that relies solely on speed inputs to infer transportation modes from dense smartphone GPS trajectories. In benchmark experiments, SpeedTransformer outperformed traditional deep learning models, such as the Long Short-Term Memory (LSTM) network. Moreover, the model demonstrated strong flexibility in transfer learning, achieving high accuracy across geographical regions after fine-tuning with small datasets. Finally, we deployed the model in a real-world experiment, where it consistently outperformed baseline models under complex built environments and high data uncertainty. These findings suggest that Transformer architectures, when combined with dense GPS trajectories, hold substantial potential for advancing transportation mode detection and broader mobility-related research. |
| title | Detecting Transportation Mode Using Dense Smartphone GPS Trajectories and Transformer Models |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2603.00340 |