WAVEFORM ADAPTIVE PROCESSING FOR LFS SONAR

We present a waveform adaptive approach to processing active sonar signals whose waveforms are distorted by propagation in shallow water dense multipath environment. Our approach is based on parametric modeling of the received signals, assuming a known signal structure with unknown parameters. The optimal processor consists of a bank of matched filters "tuned" to all possible values of the signal parameters. Direction implementation of this processor is computationally prohibitive. However, a practical computationally efficient implementation is proposed using the Polynomial Phase Transform (PPT). The PPT is a novel and powerful nonlinear transform for the analysis of highly non-stationary signals. This transform have been used very successfully for estimating the parameters of constant-modulus signals of the type commonly used in radar and sonar systems. One of its remarkable properties is that the PPT converts broadband signals whose phase is a polynomial function of time (Linear and Quadratic FM, for example) signal into a single tone! The objective of the proposed project is to develop and test the waveform adaptive processor and its PPT-based implementation. In Phase I the potential of the proposed algorithm will be demonstrated by a combination of analysis and tests on synthetic data, and some recorded data. An extensive testing on real acoustic data will take place in Phase II.