SNR‐dependent drone classification using convolutional neural networks

Abstract

Radar sensing offers a method of achieving 24-h all-weather drone surveillance, but in order to be maximally effective, systems need to be able to discriminate between birds and drones. This work examines drone-bird classification performance as a function of signal to noise ratio (SNR). Classification at low SNR values is necessary in order to classify drones with a small radar cross-section (RCS), as well as to facilitate reliable classification at longer ranges. To investigate the relationship between classification performance and SNR, Gaussian noise is added to an experimentally obtained dataset of radar spectrograms. Classification is performed by convolutional neural networks (CNNs). It is shown that for the data available classification accuracy drops with falling SNR, as might be expected for any given CNN. The degree to which performance degrades with reduced SNR is presented. It is further shown that simpler network architectures are more robust to noise. Finally, it is demonstrated that data augmentation can be used as a means of enhancing classification accuracy at lower SNR values. Bayesian optimisation is used to find the optimal augmentation hyperparameters and overall, classification accuracies of 92% are achieved at low SNR.This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Convolutional Neural Networks for Classification of Drones Using Radars

“…One would expect that classification performance decreases with the SNR because the target becomes less clear. Dale et al showed this to be true when distinguishing drones from birds [32]. As seen in Equation (1), the SNR is directly related to the RCS and so consequently tends to be small for drones.…”

A Review of Automatic Classification of Drones Using Radar: Key Considerations, Performance Evaluation, and Prospects

“…and/or employ classical spectral analysis tools (e.g., cepstrum). Recent ATC algorithms on the other hand are increasingly data-driven and leverage advances in Machine Learning (ML), such as Deep Neural Networks (DNNs), to achieve impressive classification performance, see [14]- [21].…”

“…Nevertheless, bird wings motion can result in intricate micro-Doppler-type spectral features for appropriately short radar wavelengths [24]; they are notably distinctive from those originating from mUAS blades rotating up to several thousand times a minute (i.e., drone's propellers move at a very different speed to a bird's wings). On detecting micro-Doppler signatures for ATC, convolutional Neural Networks (CNNs) have shown great promise [14]- [16], [18], [19], [21]. Their input can be Doppler spectrograms from multiple radar frames/scans (e.g.…”

Development of a networked photonic‐enabled staring radar testbed for urban surveillance

“…As expected, the longer-range data have lower SNR with a marked reduction in the number of HElicopter Rotor Modulation (HERM) lines that are visible. Previous work has reported the impact of SNR on classification performance [11]. Data from the UoB testbed are progressing the work on developing robust classifiers for challenging realistic conditions and an example of this is detailed in Table 3 from recently published results using Convolutional Neural Networks (CNN) for classifying drones and birds [12].…”