Detecting Transportation Mode Using Dense Smartphone GPS Trajectories and Transformer Models

2026-02-27

AI Summary

Accurately identifying transportation modes from dense smartphone GPS trajectories remains a key challenge in GeoAI and transportation research. This work proposes SpeedTransformer, a novel model that leverages only raw speed time series as input and employs a Transformer architecture for transportation mode recognition. By discarding complex features such as spatial coordinates and relying solely on temporal speed information, the approach maintains model simplicity while demonstrating strong cross-regional transferability and robustness to high-noise data. Experimental results show that SpeedTransformer significantly outperforms conventional models like LSTM across multiple benchmark datasets and achieves high accuracy and practicality when deployed in real-world urban environments.

Technology Category

• Planning, Routing, and Scheduling: Activity and Plan Recognition
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Application Domains: Transportation

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Security and Privacy: Data transparency and provenance

Abstract

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.

Problem

• transportation mode detection
• GPS trajectories
• GeoAI
• smartphone sensing
• mobility analysis

Innovation

• SpeedTransformer
• transportation mode detection
• dense GPS trajectories
• Transformer model
• transfer learning