An Automatic Trigged Rectangular Rapid Flashing Beacons (RRFB) System Using the Roadside LiDAR Sensor

Abstract

The existing rectangular rapid flashing beacons (RRFB) could not be automatically trigged with acceptable accuracy. This paper introduces a strategy to develop an automatic trigged RRFB system using the roadside Light Detection and Ranging (LiDAR). The trajectories of the road users can be extracted in real-time with the proposed LiDAR data processing algorithms. The pedestrian/bicycle crossing intention was predicted with Naïve Bayes (NB) classifier. Different detection zones were provided to detect the pedestrian/bicycle crossing movements in real-time. The updated RRFB system was implemented at one intersection in Henderson, Nevada, United States. The results of the field test showed that the overall false rate of the updated RRFB system is very low (0.68%). The performance of the proposed method is superior compared to the state-of-the-art method. The updated RRFB system can greatly improve pedestrian safety at unsignalized intersections. INDEX TERMS Roadside LiDAR, vehicle-pedestrian conflicts, object tracking, rectangular rapid flashing beacons, pedestrian safety.