Abstract:
Aiming at the problem that the time width and frequency modulation slope of the non-uniform smeared spectrum(SMSP) interference sub-pulse change randomly and the traditional anti-interference method fails, an interference suppression method based on Rayleigh-CFAR time-frequency filtering is proposed. The received signal is mapped to the time-frequency plane by short-time Fourier transform(STFT) based on the characteristics that the non-uniform SMSP interference presents multiple linear frequency modulation trajectories with different slopes in the time-frequency domain. Based on the statistical characteristics that the amplitude of complex Gaussian white noise obeys Rayleigh distribution after STFT, the median estimation is used to estimate the noise floor robustly, and the CFAR detection threshold is constructed adaptively by frequency to identify the time-frequency unit dominated by zero interference. After the time-domain signal is reconstructed by inverse STFT, the target is restored by matched filtering. The simulation results show that the proposed adaptive threshold can still maintain a constant false alarm rate under the condition of noise power mismatch. At the same time, the method does not need any prior parameters, and effectively suppresses the non-uniform SMSP interference in a wide range of interference-to-signal ratio. The detection performance is close to the interference-free benchmark, which verifies the effectiveness and robustness of the algorithm in strong interference environment.