Target echo enhancement of HFM signals based on non-negative matrix factorization
Hyperbolic frequency modulation(HFM)signals are Doppler-insensitive signals commonly used in active sonar.In practice,target echoes are often weak owing to channel variations and reverberation effects.To address the issue of detection performance degradation in active sonar for underwater vehicles,this paper proposes a target echo enhancement method for HFM signals based on non-negative matrix factorization.This method separately de-fines base matrices for target echo and interference signals.It uses the time-frequency diagram of transmit signals as the frequency information matrix for the target echo and imposes constraints on the time information matrix of the target echo.The enhancement in the target echo is achieved through iterative processing,resulting in a matching echo detection with higher gain.Numerical simulations and sea test results show that the method improves the peak-to-average ratio by more than 2 dB compared with traditional matched filtering methods,thereby enhancing echo de-tection outcomes.
active sonarunderwater detectionreverberationnoisenon-negative matrix factorizationhyperbolic frequency modulation(HFM)signalecho enhancementtarget detection