TrajDLM: Topology-Aware Block Diffusion Language Model for Trajectory Generation
2026-05-11 • Machine Learning
Machine Learning
AI summaryⓘ
The authors developed TrajDLM, a new method for creating synthetic GPS paths that stick closely to actual road networks while being faster than earlier methods. Instead of working with continuous locations, their model uses sequences of road segments and a block diffusion process to efficiently generate realistic routes. TrajDLM also uses road network information to keep the generated paths logical and accurate. Tested on data from multiple cities, it performed well in matching real paths and was also able to work on unfamiliar transportation types without extra training.
GPS trajectoriessynthetic data generationroad network topologyblock diffusion modelsdiscrete trajectory modelingtopology-aware embeddingsdenoisingautoregressive decodingzero-shot transfertransportation modes
Authors
Wilson Wongso, Lihuan Li, Arian Prabowo, Xiachong Lin, Baiyu Chen, Hao Xue, Flora D. Salim
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/