Unifying Lane-Level Traffic Prediction from a Graph Structural Perspective: Benchmark and Baseline

2024-03-22

arXiv.org

AI Summary

Lane-level traffic forecasting has long suffered from fragmented research due to inconsistent evaluation protocols, scarce public datasets, and limited code availability. To address this, we propose the first standardized lane-level forecasting benchmark, featuring a unified graph-structured modeling framework with rigorously defined topology construction rules and task paradigms. We release three large-scale, real-world datasets—including fine-grained V2X traffic observations—and open-source the complete implementation and a fair, reproducible evaluation pipeline. Furthermore, we design GraphMLP, a lightweight baseline that synergistically integrates graph neural networks (GNNs) for spatial dependency modeling, multilayer perceptrons (MLPs) for efficient inference, and lane-specific representation learning. Extensive experiments demonstrate that GraphMLP consistently outperforms state-of-the-art methods in prediction accuracy, inference speed, and cross-scenario generalization. This work establishes a foundation for systematic, reproducible advancement in lane-level traffic forecasting.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Planning, Routing, and Scheduling: Planning with Language Models
• Computer Vision: Large Vision Models

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Traffic prediction has long been a focal and pivotal area in research, witnessing both significant strides from city-level to road-level predictions in recent years. With the advancement of Vehicle-to-Everything (V2X) technologies, autonomous driving, and large-scale models in the traffic domain, lane-level traffic prediction has emerged as an indispensable direction. However, further progress in this field is hindered by the absence of comprehensive and unified evaluation standards, coupled with limited public availability of data and code. This paper extensively analyzes and categorizes existing research in lane-level traffic prediction, establishes a unified spatial topology structure and prediction tasks, and introduces a simple baseline model, GraphMLP, based on graph structure and MLP networks. We have replicated codes not publicly available in existing studies and, based on this, thoroughly and fairly assessed various models in terms of effectiveness, efficiency, and applicability, providing insights for practical applications. Additionally, we have released three new datasets and corresponding codes to accelerate progress in this field, all of which can be found on https://github.com/ShuhaoLii/TITS24LaneLevel-Traffic-Benchmark.

Problem

• Lack of unified standards for lane-level traffic prediction
• Limited public availability of lane-level traffic data and code
• Need for consistent evaluation across datasets and models

Innovation

• Unified spatial topology structure for lane-level prediction
• GraphMLP model combining graph structure and MLP
• Public datasets covering regular and irregular lanes