MovSemCL: Movement-Semantics Contrastive Learning for Trajectory Similarity (Extension)

TL;DR

MovSemCL introduces a movement-semantics contrastive learning framework that effectively captures hierarchical trajectory semantics, reduces computational costs, and employs physically plausible augmentations, significantly improving trajectory similarity tasks.

Contribution

This work presents MovSemCL, a novel contrastive learning approach that models trajectory semantics hierarchically and efficiently, with a curvature-guided augmentation strategy to enhance similarity computation.

Findings

Outperforms state-of-the-art methods in similarity search accuracy.

Reduces inference latency by up to 43.4%.

Achieves up to 20.3% improvement in heuristic approximation.

Abstract

Trajectory similarity computation is fundamental functionality that is used for, e.g., clustering, prediction, and anomaly detection. However, existing learning-based methods exhibit three key limitations: (1) insufficient modeling of trajectory semantics and hierarchy, lacking both movement dynamics extraction and multi-scale structural representation; (2) high computational costs due to point-wise encoding; and (3) use of physically implausible augmentations that distort trajectory semantics. To address these issues, we propose MovSemCL, a movement-semantics contrastive learning framework for trajectory similarity computation. MovSemCL first transforms raw GPS trajectories into movement-semantics features and then segments them into patches. Next, MovSemCL employs intra- and inter-patch attentions to encode local as well as global trajectory patterns, enabling efficient hierarchical representation and reducing computational costs. Moreover, MovSemCL includes a curvature-guided augmentation strategy that preserves informative segments (e.g., turns and intersections) and masks redundant ones, generating physically plausible augmented views. Experiments on real-world datasets show that MovSemCL is capable of outperforming state-of-the-art methods, achieving mean ranks close to the ideal value of 1 at similarity search tasks and improvements by up to 20.3% at heuristic approximation, while reducing inference latency by up to 43.4%.