TrajMamba: An Efficient and Semantic-rich Vehicle Trajectory Pre-training Model

TL;DR

TrajMamba introduces a semantic-rich yet efficient pipeline for vehicle trajectory representation. It jointly models GPS and road perspectives with a Traj-Mamba Encoder, while Travel Purpose-aware Pre-training injects travel semantics through road/POI textual views, and Knowledge Distinction Pre-training compresses trajectories via a learnable mask. The approach yields consistent gains on Destination Prediction, Arrival Time Estimation, and Similar Trajectory Search, and demonstrates favorable model efficiency due to linear-time inference and effective pre-training. This work enables more accurate, scalable ITS applications by combining robust travel semantics with efficient trajectory compression.

Abstract

Vehicle GPS trajectories record how vehicles move over time, storing valuable travel semantics, including movement patterns and travel purposes. Learning travel semantics effectively and efficiently is crucial for real-world applications of trajectory data, which is hindered by two major challenges. First, travel purposes are tied to the functions of the roads and points-of-interest (POIs) involved in a trip. Such information is encoded in textual addresses and descriptions and introduces heavy computational burden to modeling. Second, real-world trajectories often contain redundant points, which harm both computational efficiency and trajectory embedding quality. To address these challenges, we propose TrajMamba, a novel approach for efficient and semantically rich vehicle trajectory learning. TrajMamba introduces a Traj-Mamba Encoder that captures movement patterns by jointly modeling both GPS and road perspectives of trajectories, enabling robust representations of continuous travel behaviors. It also incorporates a Travel Purpose-aware Pre-training procedure to integrate travel purposes into the learned embeddings without introducing extra overhead to embedding calculation. To reduce redundancy in trajectories, TrajMamba features a Knowledge Distillation Pre-training scheme to identify key trajectory points through a learnable mask generator and obtain effective compressed trajectory embeddings. Extensive experiments on two real-world datasets and three downstream tasks show that TrajMamba outperforms state-of-the-art baselines in both efficiency and accuracy.

Figures (6)

• Figure 1: A trajectory of commuting to work.
• Figure 2: The framework of TrajMamba.
• Figure 3: Structure of Traj-Mamba blocks.
• Figure 4: Effectiveness of hyper-parameters validated on Chengdu dataset.
• Figure 5: Scalability of fine-tuning on Chengdu dataset.