E-Scooter Rider detection and classification in dense urban environments

Accurate detection and classification of vulnerable road users is a safety critical requirement for the deployment autonomous vehicles in heterogeneous traffic. Although similar physical appearance to pedestrians, e-scooter riders follow distinctly different characteristics movement can reach speeds up 45kmph. The challenge detecting exacerbated urban environments where frequency partial occlusion increased as navigate between vehicles, traffic infrastructure other users. This lead non-detection or mis-classification providing inaccurate information accident mitigation path planning vehicle applications. research introduces novel benchmark partially occluded rider facilitate objective characterization models. A novel, occlusion-aware method presented that achieves 15.93% improvement performance over current state art.