TF-Lane: Traffic Flow Module for Robust Lane Perception

TL;DR

This paper introduces TF-Lane, a traffic flow module that enhances lane perception in autonomous driving by leveraging real-time traffic flow data, especially in occluded or missing lane scenarios, without high costs.

Contribution

The paper presents a novel traffic flow-aware module that improves lane perception robustness by integrating real-time traffic data with existing algorithms, validated on multiple models and datasets.

Findings

Up to +4.1% mAP improvement on Nuscenes dataset.

Consistent performance gains across four mainstream models.

Effective in occluded or lane-missing scenarios.

Abstract

Autonomous driving systems require robust lane perception capabilities, yet existing vision-based detection methods suffer significant performance degradation when visual sensors provide insufficient cues, such as in occluded or lane-missing scenarios. While some approaches incorporate high-definition maps as supplementary information, these solutions face challenges of high subscription costs and limited real-time performance. To address these limitations, we explore an innovative information source: traffic flow, which offers real-time capabilities without additional costs. This paper proposes a TrafficFlow-aware Lane perception Module (TFM) that effectively extracts real-time traffic flow features and seamlessly integrates them with existing lane perception algorithms. This solution originated from real-world autonomous driving conditions and was subsequently validated on open-source algorithms and datasets. Extensive experiments on four mainstream models and two public datasets (Nuscenes and OpenLaneV2) using standard evaluation metrics show that TFM consistently improves performance, achieving up to +4.1% mAP gain on the Nuscenes dataset.