Relative pose estimation for cooperative vehicles and tracking of multiple dynamic objects

A cooperative perception (CP) between two vehicles to work correctly requires relative positions between two vehicles to be established through a robust and accurate estimation method. This study focuses on relative pose estimation between two vehicles for cooperative perception-based autonomous driving through indirect methods. The indirect method utilizes the positioning data sharing from the Inertial Navigation System(INS) of each vehicle. The relative pose estimation is highly effective in cooperative perception where uncertainties are often involved due to full or partial occlusions caused by dynamic or moving objects such as Vulnerable road users (VRUs) to avoid hazardous situations. The objective is to enhance each autonomous vehicle's perception capability beyond its own field of view through cooperation with other surrounding vehicles. Hence, in this study, the relative pose estimation between two vehicles in a dynamic situation is presented for cooperative perception-based automated driving system (ADS). The study also presents the detection and tracking results of multiple dynamic objects such as bicyclists, pedestrians, and other vehicles from the scene to avoid accidents or collisions during uncertain or critical situations by sharing their perception information.