Comparative Analysis of Radar-Cross-Section-Based UAV Recognition Techniques

This work investigates the problem of unmanned aerial vehicle (UAV) recognition using their radar cross section (RCS) signature. The RCS of six commercial UAVs is measured at 15 and 25 GHz in an anechoic chamber for both vertical-vertical (VV) polarization and horizontal-horizontal (HH) polarization. The RCS signatures are used to train 15 different recognition algorithms, each belonging to one of three different categories: statistical learning (SL), machine learning (ML), and deep learning (DL). The study shows that, while the average accuracy of all the algorithms increases with the signal-to-noise ratio (SNR), the ML algorithm achieved better accuracy than the SL and DL algorithms. For example, the classification tree ML achieves an accuracy of 98.66% at 3-dB SNR using the 15-GHz VV-polarized RCS test data from the UAVs. We investigate the recognition accuracy using the Monte Carlo analysis with the aid of boxplots, confusion matrices, and classification plots. On average, the accuracy of the classification tree ML model performed better than the other algorithms, followed by Peter Swerling's statistical models and the discriminant analysis ML model. In general, the accuracy of the ML and SL algorithms outperformed the DL algorithms (Squeezenet, Googlenet, Nasnet, and Resnet 101) considered in the study. Furthermore, the computational time of each algorithm is analyzed. The study concludes that, while the SL algorithms achieved good recognition accuracy, the computational time was relatively long when compared to the ML and DL algorithms. Also, the study shows that the classification tree achieved the fastest average recognition time of about 0.46 ms.