A Novel Energy Aware Secure Internet of Drones Design: ESIoD

Abstract

Unmanned aerial vehicles (UAVs), or drones, are emerging as a promising technology for a variety of monitoring and surveillance-based applications. Smart UAVs are not limited only to image capturing, but also to real-time decision making using artificial intelligence. Moreover, it is important to consider the data security of captured images. In this paper, we propose a novel Energy-aware Secure Internet of Drone (ESIoD) architecture. A crucial research problem addressed by this work is how to accomplish faster onboard processing and reduce battery usage for a UAV to prolong the flight time while retaining data security of UAV captured images. Specifically, drone-captured real-time images are encrypted using either AES or RSA algorithms and offloaded by the onboard computer to a cloud server for the processing of cognitive actions using both a standard Haar cascade classifier and an advanced faster R-CNN classifier. The focus of this study is to conserve the drone battery life by secure computational offloading to optimize drone flight time. Two sets of experiments were performed using drone-captured sample images and videos. Results show that the ESIoD architecture can conserve 80% onboard processing time and 3X drone battery charge usage as compared to conventional real-time onboard processing for the considered application.