An improved sensing method using radio frequency detection

Easy-to-pilot unmanned aerial vehicles (UAVs) are now readily available off the shelf. This has created the problem of micro unmanned aerial vehicle (MUAV) in private or sensitive areas which can represent a personal or public threat. An improved MUAV detection method is proposed in this paper using artificial neural network (ANN) based feature extraction and reconstruction. The clutters form static/non-static objects in the collected radio frequency (RF) signals are suppressed via employing the background estimate method and suppressing the linear trend. Then principal component of the RF signals are extracted based on the singular value decomposition (SVD) method. The higher order cumulant (HOC) algorithm is utilized to improve the signal to noise ratio (SNR) of the RF signals, which can make Gaussian noise prone to zero. Hilbert spectrums of the analyzed features are considered to determine if one MUAV is present in the detection area using ANN. Finally, the region of interest (ROI) containing RF signals is defined to estimate the azimuth and first frequency of MUAV. Detection results in real-life scenarios are obtained which show the effectiveness of the proposed technique in detecting MUAV. (C) 2019 Elsevier B.V. All rights reserved.