Complex Flexible Analytic Wavelet Transform for UAV State Identification Using RF Signal

The time-frequency analysis is a highly suited technique for nonstationary signal analysis which studies a signal in both time and frequency domains simultaneously. The combination of real-time signals of two systems hold quadrature property and become complex in nature. In such cases, information is distinct in positive and negative frequency ranges and can be utilized for signal analysis. In this article, the flexible analytic wavelet transform (FAWT) is extended to decompose a complex signal in positive and negative frequency ranges. The Hilbert transform (HT) is applied to formulate the time-frequency representation with positive and negative frequency ranges without using ideal band-pass filter. Moreover, a genetic algorithm-based method is developed for the parameter optimization of FAWT with respect to minimization of bandwidth in the low-pass frequency of the last level. The proposed method is compared with the existing method and extended for unmanned aerial vehicles (UAV) state identification using radio frequency (RF) signal intercepted in clean, blue-tooth, Wi-Fi (WIFI), and both types of noisy environment. The complex RF signal is decomposed into positive and negative frequency components which are utilized for statistical features computation and classification. The UAV state identification system employed two stage identifications, initially for UAV type identification followed by state identification. The developed method gives promising results for UAV type and state identification which is useful for UAV surveillance system development.