Efficient Face Recognition via Multi-UAV-Edge Collaboration in UAV Delivery Service

Abstract

In recent years, smart unmanned aerial vehicle (UAV) delivery has become a promising solution to solve the last-mile delivery problem in smart logistics. In a smart UAV delivery system, the accurate identification of the target consignee is a critical task. Face recognition as an important AI application for user authentication, existing cloud computing based face recognition methods often face several critical issues such as massive data transmission to the cloud, slow response time, and the concern over user data privacy. To address these issues, in this paper, we propose a multi-UAV-edge collaborative framework for accurate, efficient and secure face recognition in the edge computing environment. Specifically, the feature values are extracted from the target consignee's face images and then stored on the permissioned blockchain node. Second, a parallel task processing framework for face recognition on edge servers is proposed. Finally, the results of face recognition are broadcasted to all UAVs so that they are working collaboratively to identify their target consignee. Simulation experiments based on a real-world smart UAV delivery system demonstrate the effectiveness of the proposed multi-UAV-edge collaborative framework.