UAV-Assisted Blind Area Pedestrian Detection via Terminal-Edge-Cloud Cooperation in VANETs

Qisen Zhang, Kai Liu, Luyao Jiang, Chengliang Zhong, Liang Feng, Hao Zhang, Ke Xiao

Abstract: The collaboration of UAV (Unmanned Aerial Vehicle) and VANETs (Vehicular Ad-hoc Networks) on enabling future intelligent transportation systems has received great attention in both academia and industries. In this paper, we make the first effort on presenting a terminal-edge-cloud cooperation architecture for UAV assisted blind area pedestrian detection in VANETs. On this basis, a task offloading algorithm is proposed, which offloads the computation task adaptively based on heterogeneous computation, communication and storage capacities of UAVs, vehicles, roadside infrastructures and the cloud server, aiming at minimizing the total service delay by striking a best balance between task computation delay and data transmission delay. Then, we implement the system prototype and conduct hardware-in-the-loop experiments to verify the effectiveness of the proposed algorithm. Finally, we test the system performance in realistic VANETs environments and demonstrate the feasibility of the proposed architecture and solution.