Being More Lightweight and Practical: Mini-sized Contrastive Learning Pre-trained Models for Fine-grained Traffic Task

Fine-grained traffic prediction is critically important for mitigating traffic congestion in key urban areas and for providing lane-change guidance in autonomous vehicles and navigation systems. However, task-specific models are not efficient enough, city-scale pre-trained models often overlook fine-grained requirements, and the demand for extensive computational resources hinders practical deployment. To address this issue, we developed a lightweight pre-training framework, MiniTraffic. This framework leverages abundant road-level data to address lane-level data scarcity through a frequency domain stability augmentation module and captures road-lane correlations via contrastive clustering to construct small-scale graph structures, significantly reducing model parameters. Fine-tuning with minimal target data provides a unified and efficient solution for fine-grained traffic prediction. In multi-granularity traffic prediction tasks across six fine-grained datasets, MiniTraffic demonstrated superior performance compared to existing baselines.

Modern navigation and self-driving systems increasingly need lane-level traffic predictions, not just for the road as a whole. But lane-level traffic sensors are rare, while road-wide data is abundant, and existing models are either too narrow or too large to run on real-world devices. We developed MiniTraffic, a compact model that learns general traffic patterns from abundant road-wide data and adapts to lane-level prediction using only a small amount of lane data. MiniTraffic outperforms much larger models while staying small enough for edge deployment, supporting smarter signal control and lane-change guidance.