Signal processing methods for active synthetic aperture sonar

In the autumn 2001 FOI carried out a field experiment at the Älvsnabben test site in the southern Stockholm archipelago. The aim of the experiment was to test Synthetic Aperture Sonar (SAS) algorithms developed at FOI during the last two years. These algorithms had previously been used to analyse data from an earlier experiment at the Djupviken test site with good results. The same equipment was used in both experiments, but with minor modifications to the transmitter. Studies of the recorded data revealed several shortcomings in the experimental equipment and experimental mistakes which made the analysis very complicated. Most problems stemmed from the lack of a proper navigation instrument on the Remotely Operated Vehicle (ROV) carrying the sonar. There were also incompatibility problems with some of the rented equipment. Even though none of the runs were good some analysis was still possible. Autopositioning algorithms were extended to exploit Fast Factored Back Projection (FFBP) imaging. FFBP was originally developed at FOI for wideband Synthetic Aperture Radar (SAR). We give a status report from the SAS research, as well as an introduction to SAS theory. In combination with the previous Djupviken experiment, our conclusions from the Älvsnabben experiment are that the algorithms work well, but that the sonar and navigation systems must be upgraded. The report concludes with suggestions towards a system specification.