Low SNR radar signal detection using the continuous wavelet transform (CWT) and a Morlet wavelet

A continuous wavelet transform (CWT) system for low SNR radar target detection is proposed in this paper. The radar transmission signal is a pulsed complex exponential. The proposed method automatically determines the optimal scale parameter for CWT analysis by finding the peak in the return signal CWT power computed across translations. The contributions of this paper are an algorithm for determining the optimal scale range for integration of CWT data, a denoising algorithm based on analysis in a small neighborhood around the optimal scale, and analysis of computational and memory requirements. Results are analyzed for simulated radar targets corrupted by white Gaussian noise and compared to a more traditional matched filter (MF). Theoretical and simulation results show that the average CWT and MF process gain (PG) is 23.4 dB and 21.2 dB, respectively, with no Doppler shift, where the theoretical CWT PG is 24.1 dB. For Doppler shifted signals, the mean PG was 23.53 and 13.84 dB for CWT and MF, respectively. The proposed method achieves 3 to 10 dB on average higher PG than the MF method.