Understanding Real-World Brake Activity: A Key to Assessing Non-Tailpipe Emission Sources

Understanding and mitigating non-tailpipe emissions, particularly those from braking activities, is crucial for improving air quality and public health. This study investigates the impact of vehicle brake activity on mobile-source particulate matter (PM) emissions at signalized intersections using a multi-modal data collection approach. The research integrates LiDAR-based trajectory detection, roadside camera monitoring, drone surveillance, and in-vehicle sensor data to analyze braking behaviors and their associated emissions. A hybrid physical-machine learning model was developed to detect braking events and estimate brake-induced PM emissions. The methodology involves LiDAR-based vehicle tracking, trajectory refinement, brake activity detection, and brake emission estimation. By leveraging deep learning algorithms for multi-object tracking and data fusion techniques, the study provides a robust framework for real-time brake activity monitoring. The study further incorporates an empirical emissions model, which is calibrated using laboratory dynamometer tests, to estimate brake wear particle emissions based on kinetic energy dissipation. The results demonstrate that braking intensity and frequency significantly influence PM emissions, with variations observed across different brake pad materials. Real-world validation using a probe vehicle dataset confirms the accuracy of the model, achieving high precision (97%) and recall (100%) in braking event detection. The brake emissions estimation model effectively quantifies PM10 emissions, with calibrated emission factors highlighting material-specific variations. Findings from this research underscore the importance of accounting for non-tailpipe emissions in urban air quality management and regulatory policies. The proposed methodology provides a scalable, non-intrusive solution for monitoring braking-induced emissions in real-world traffic conditions. Future research should explore the integration of regenerative braking impacts, powertrain differentiation, and further refinements in sensor fusion techniques to enhance emission inventory accuracy and inform transportation policy development. View the NCST Project Webpage