Machine learning-enabled UAV Networking

We leverage machine learning to improve the quality of service for the 5G-enabled Unmanned Aerial Vehicle (UAV) networking which is important to increase the scale of UAV swarms. 

5G-enabled Optimal Bi-Throughput For UAS Swarm Networking

Knowledge Distillation based GPS Spoofing Detection for Small UAS

As a core component of small Unmanned Aerial Vehicle (UAV), GPS is playing a critical role to provide localization for UAV navigation. UAV is an important factor in the large scale deployment of the Internet of Things (IoT) and Cyber-Physical Systems (CPS). However, GPS is vulnerable to spoofing attacks that can mislead UAV to fly to a sensitive area and make threats to public safety and private security. The conventional spoofing detection methods need too much overhead which stops an efficient classifier to work in a computation-constrained UAV and provides an efficient response to the attacks. In this paper, we proposed a novel approach to implement a lightweight classifier in UAV system so that GPS spoofing attacks can be detected from a long distance. With Recurrent Neural Networks (RNN), we proposed a novel classifier on the ground control stations and then we transferred it into the control system of UAV with knowledge distillation in a compact size. The experimental results show that our lightweight classifier runs in UAV systems reliably and it can achieve a good accuracy of GPS spoofing detection.