Reverberation Rejection With Active Line Towed Array

Detection in the presence of reverberation is often difficult in active sonar, due to the strong interference from the ocean. In recent years, various kinds of theories and ways are proposed worldwide to reject reverberation, but the result is not very good.In this paper, the "Principal Component Inverse" (PCI) algorithm is selected to detect targets in reverberation environment by the comparison of a lot of data obtained from an active towed line array. It is a simplified model which is more deterministic: reverberation is considered as a sum of undesirable echoes. The singular value decomposition of matrix and its physical meanings are introduced, which indicates that the singular value can be divided into two subspace vectors and estimate ranks of the observed matrix, thus forming the mathematical and physical basis for the PCI algorithm. It shows that the singular value decomposition is valid in distinguishing target signal from reverberation using PCI. The PCI algorithm is also extended to be used in spatio-temporal and spectra domain, which shows that high performance can be obtained in distinguishing between targets and reverberation. The iterative PCI algorithm is tried to reject reverberation.Finally, the comparison of the adaptability of the PCI algorithm in different S/R (signal to reverberation ratio) shows that the PCI algorithm has different gains in different S/R. A best primitive S/R is presented. Hence, the detection performance can be improved in small S/R using the PCI algorithm.