The Application of Go-RA Decomposition Algorithm in Shallow Water Reverberation

Abstract Shallow bottom reverberation is a primary factor that limits the performance of active detection, and the separation of reverberation, target, and noise components from received signals is a hot topic in underwater acoustic signal processing. This paper proposes an improved Go-RA decomposition method based on randomized singular value decomposition (RA-SVD). The method can successfully separate the reverberation background and strong target echoes at the simulation level. Moreover, compared with the original Go algorithm, the proposed Go-RA method significantly enhances computational efficiency without compromising the accuracy of the separation results. This method has broad application prospects, applicable not only in seabed detection, extraction of reverberation waveguide invariants, and data-driven reverberation suppression but also in the separation of strong target signals.