Ensemble Deep Learning for Sustainable Multimodal UAV Classification

Unmanned aerial vehicles (UAVs) have increasingly shown to be useful in civilian applications (such as agriculture, public safety, surveillance) and mission critical military applications. Despite the growth in popularity and applications, UAVs have also been used for malicious purposes. In such instances, their timely detection and identification has garnished rising interest from government, industry and academia. While much work has been done for detecting UAVs, there still exist limitations related to the impact of extreme environmental conditions and big dataset requirements. This paper proposes a novel ensemble deep learning framework that has hybrid synthetic and deep features to detect unauthorized or malicious UAVs by using acoustic, image/video and wireless radio frequency (RF) signals for robust UAV detection and classification. We present the performance evaluation of the proposed approach using numerical results obtained from experiments using acoustic, image/video and wireless RF signals. The proposed approach outperforms the existing related approaches for detecting malicious UAVs.

I. Introduction

Advances in inexpensive and widely available commercial and consumer grade unmanned aerial vehicles (UAVs) or “drones” have expanded their overall use. These rapid developments and deployment of wide spread applications of UAVs have led to an increased need to address various concerns related to public safety and privacy. It is critical detect and identify malicious UAVs flying in unauthorized or restricted areas [1]. Malicious UAVs can be considered as those that are outfitted with explosive payloads or that are used to collect data in restricted territory. Low altitude flights may allow them to operate in restricted areas without triggering traditional air-space security measures. Restricted air-spaces can be understood as areas (land or water/ocean) above which unauthorized UAVs are not permitted under certain conditions. To illustrate the overall concept, a typical system model is depicted in Fig. 1 where we can see UAVs, UAV-restricted area and UAV-monitoring area. Fig. 1.

System model for detecting malicious UAVs with UAV restricted zone and monitoring area.