LSM: A Comprehensive Metric for Assessing the Safety of Lane Detection Systems in Autonomous Driving
🤖 AI Summary
Existing safety evaluations for lane detection rely excessively on accuracy metrics (e.g., mAP) and lack interpretable, functionally safe measures that integrate road semantics (e.g., type, width), dynamic operational conditions (e.g., vehicle speed, detection range), and root-cause analysis of failures.
Method: We propose the Lane Detection Safety Metric (LSM), the first functionally safe evaluation standard for lane detection. LSM unifies road structural semantics, real-time operational constraints, and causal analysis of missed detections into a scenario-aware, causality-grounded safety scoring function, implemented via offline quantification in a virtual simulation platform.
Contribution/Results: Experiments across diverse complex virtual driving scenarios demonstrate that LSM effectively discriminates the safety performance of different algorithms. Its safety rankings strongly correlate with actual risk exposure—outperforming conventional performance metrics in predictive validity and interpretability.
Application Category
📝 Abstract
Comprehensive perception of the vehicle’s environment and correct interpretation of the environment are crucial for the safe operation of autonomous vehicles. The perception of surrounding objects is the main component for further tasks such as trajectory planning. However, safe trajectory planning requires not only object detection, but also the detection of drivable areas and lane corridors. While first approaches consider an advanced safety evaluation of object detection, the evaluation of lane detection still lacks sufficient safety metrics. For assessing safety, additional factors such as the semantics of the scene with road type and road width, the detection range as well as the potential causes of missing detections, incorporated by vehicle speed, should be considered for the evaluation of lane detection. Therefore, we propose the Lane Safety Metric (LSM), which takes these factors into account in order to evaluate the safety of lane detection systems by determining an easily interpretable safety score. We evaluate our offline safety metric on various virtual scenarios using different lane detection approaches and compare it with state-of-the-art performance metrics.
Problem
Research questions and friction points this paper is trying to address.
Evaluates lane detection safety in autonomous driving
Considers road semantics, detection range, and missing causes
Proposes LSM metric for interpretable safety scoring
Innovation
Methods, ideas, or system contributions that make the work stand out.
Proposes Lane Safety Metric (LSM) for lane detection safety
Incorporates scene semantics, detection range, and missing causes
Evaluates safety offline using virtual scenarios and comparisons
🔎 Similar Papers
No similar papers found.
