RESUMEN
This letter proposes a reverberation suppression method to improve the target detection probability in reverberation environment induced by a linear frequency modulation signal. Using the singular value decomposition and an assumption of local spatial stationarity, the proposed algorithm suppresses reverberation utilizing the difference of singular values between the reference beam and the current beam of interest. Several experiments using real oceanic reverberation data justify the validity of the proposed algorithm.
Asunto(s)
Aire , Procesamiento de Señales Asistido por Computador , Sonido , Algoritmos , Inteligencia Artificial , Océanos y MaresRESUMEN
This letter presents an autoregressive (AR) prewhitener for linear frequency modulation (LFM) reverberation to enhance the target signal. The proposed method uses a dechirping transformation to inversely compensate the frequency chirp rate of the LFM and give the LFM reverberation a stationary frequency property in each data block. The left or right beam signal adjacent to the current beam is then used as the reference signal, and the frequency response of each data block modeled using the AR coefficients. Finally, these coefficients are used to implement the inverse filter and efficiently prewhiten the LFM reverberation of the current beam.