TrajDLM: Topology-Aware Block Diffusion Language Model for Trajectory Generation

2026-05-11

AI Summary

Existing methods for generating high-fidelity GPS trajectories struggle to balance computational efficiency with faithfulness to road network topology. This work proposes the first application of block diffusion language models to trajectory generation, modeling trajectories as sequences of discrete road segments. By integrating topology-aware embeddings from a road network graph encoder and a topology-constrained sampling strategy, the approach achieves both high generation speed and structural consistency. Evaluated on three city-scale datasets, the method significantly outperforms baseline models, particularly in local similarity metrics, and achieves up to a 2.8× speedup in generation time. Moreover, it demonstrates strong zero-shot cross-domain transferability, highlighting its generalization capability across diverse urban environments.

Abstract

Generating high-fidelity synthetic GPS trajectories is increasingly important for applications in transportation, urban planning, and what-if scenario simulation, especially as privacy concerns limit access to real-world mobility data. Existing trajectory generation models face a trade-off between efficiency and faithfulness to road network topology: continuous-space methods enable fast generation but ignore the road network, while topology-aware approaches rely on search-based autoregressive decoding that limits generation speed. We propose TrajDLM, a topology-aware trajectory generation framework based on block diffusion language models that bridges this gap. TrajDLM models trajectories as sequences of discrete road segments, combining a block diffusion backbone for efficient denoising, topology-aware embeddings from a road network encoder, and topology-constrained sampling to ensure coherent and realistic trajectories. Across three city-scale datasets, TrajDLM achieves strong performance on fine-grained local similarity metrics while being up to $2.8\times$ faster than prior work, and demonstrates strong zero-shot transfer across domains, including unseen transportation modes. These results highlight the effectiveness of block-wise discrete diffusion as a scalable approach to accurate and efficient trajectory generation. Our code is available at https://github.com/cruiseresearchgroup/TrajDLM/

Problem

• trajectory generation
• road network topology
• synthetic GPS trajectories
• generation efficiency
• topology-aware modeling

Innovation

• block diffusion
• topology-aware generation
• trajectory synthesis
• discrete diffusion language model
• road network embedding