High resolution methods for sonar

Jeffries, David John (1986) High resolution methods for sonar. University of Southampton, Doctoral Thesis.

Abstract

This thesis addresses the problem of underwater detection, using arrays of passive sonar buoys. The performance of a number of high resolution adaptive signal processing algorithms are examined. These methods are used for the detection of narrowband signals in broadband noise and the processing takes place in the frequency domain. The majority of arrays studied in this research are linear with the sensors equispaced along the array. The study incorporates both theoretical and experimental results. The analytical results are used to predict the performance of selected algorithms. Other methods are examined by extensive simulation studies, for specific signal and noise fields. Some new techniques are introduced, which attempt to use prior information more efficiently than present methods. This approach is shown to improve the performance of the well known maximum entropy method, particularly when the noise field is partially known. The more recent eigenvector methods are capable of very high resolution, but it is demonstrated that including alternative prior information can improve their performance. The new algorithms are more robust than conventional eigenvector methods and can give useful results in the presence of phase and amplitude errors, while retaining high resolution properties. Throughout the research more emphasis has been placed on producing high resolution algorithms, than on developing numerically efficient techniques.

This record has no associated files available for download.

Contributors

Download statistics