A macroscopic traffic data fusion method and practice based on highway roadside perception

This study is based on data from roadside cameras and millimeter-wave radar on expressways. It focuses on developing state estimation methods for individual vehicles, rule-based analysis methods for section-level traffic events, and a traffic congestion data fusion method using density clustering DBSCAN algorithms. The study conducted practical collection and fusion of raw roadside perception data on the Chengdu-Yibin Expressway. Based on these fusion methods, it achieved the integration of five types of macroscopic traffic data: traffic flow, traffic density, traffic speed, traffic events (such as illegal parking, wrong-way driving, speeding, etc.), and traffic congestion.