RF-Based UAV Detection and Identification Using Hierarchical Learning Approach

Abstract

Unmanned Aerial Vehicles (UAVs) are widely available in the current market to be used either for recreation as a hobby or to serve specific industrial requirements, such as agriculture and construction. However, illegitimate and criminal usage of UAVs is also on the rise which introduces their effective identification and detection as a research challenge. This paper proposes a novel machine learning-based for efficient identification and detection of UAVs. Specifically, an improved UAV identification and detection approach is presented using an ensemble learning based on the hierarchical concept, along with pre-processing and feature extraction stages for the Radio Frequency (RF) data. Filtering is applied on the RF signals in the detection approach to improve the output. This approach consists of four classifiers and they are working in a hierarchical way. The sample will pass the first classifier to check the availability of the UAV, and then it will specify the type of the detected UAV using the second classifier. The last two classifiers will handle the sample that is related to Bebop and AR to specify their mode. Evaluation of the proposed approach with publicly available dataset demonstrates better efficiency compared to existing detection systems in the literature. It has the ability to investigate whether a UAV is flying within the area or not, and it can directly identify the type of UAV and then the flight mode of the detected UAV with accuracy around 99%.

Insight of Unmanned Aerial Vehicles Accessing Ensemble Techniques

“…The amalgamation of ensemble learning with unmanned aerial vehicles has opened up immense options for people all across the world belonging to diverse domains. In this paper, a recent survey based of ensemble enabled applications has [22] Identification and detection Efficiency KNN, XGBoost XGBoost with outlier detection [23] High resolution data updation Speed volume Deep learning with semantic segmentation Semantic segmentation with RFID [24] Timely data acquisition Field scale Random forest, support vector regression, K-nearest neighbours Random forest with RFID been performed outlining the importance of enhancements in the 5G wireless network arena. Various issues popping up due to this integration and their associated low complexity solutions (both existing and proposed) are tabulated.…”

“…References Application domain Description use of ensemble with UAV [5,6] Wireless communications Received signal strength prediction, throughput optimization, improved ground to air communications [7,18] Image processing Global positioning, processing of spatial data [16] Forensics and security Gully headcut location prediction [8,11] Geomorphology [9][10][11][12][13] Agriculture Health monitoring of crop, location intelligence, band selection and validity, multispectral imagery for soil moisture and texture [22] Identification and detection Filtering of RF signals and different classifiers checking the availability and detect ability of UAV [23] Photogrammetry Improvement of extraction accuracy using deep learning semantic segmentation model applied to UAV images, segregation of building and ground objects with RGBD bands [24] Hyperspectral imagery Hyperspectral images extraction and pre-processing, spectral feature extraction and reduction, yield statistics and spectral profile depending on harvesting time…”

“…In a very recent study ensemble learning based UAV identification and detection has been figured out based on the feature extraction of radio frequency data for efficiently investigating whether UAV is flying in the designated area or not [22].…”

Drone Detection Sensor With Continuous 2.4 GHz ISM Band Coverage Based on Cost-Effective SDR Platform

“…The RF capture device can be based on high frequency oscilloscope [18], spectrum analyzer [20] or SDR platform [21]- [24]. Although both, oscilloscope and spectrum analyzer, offers high accuracy and fulfills bandwidth requirement, they are designed for rather a laboratory use.…”

“…On the other hand, solutions utilizing inexpensive SDR suffer from the limited 40 MHz bandwidth [23], allowing to observe only half of the needed spectrum in real-time. The Authors in [23], [24] used dual SDRs running simultaneously to obtain a machine learning dataset and overcome this limit. Some methods based on statistical signatures like skewness, kurtosis, and standard deviation, can be applied using a small band of the signal in the frequency domain [25].…”