Automated Broadband Acoustic Sound Velocity Profiler

Scientific Fishery Systems, Inc. (SciFish) recently completed a Phase I SBIR that used neural networks to learn the relationship between temperature and the spectra of broadband returns for a fixed location during a fixed time of the year. This approach relaxes the assumption of the consistency of backscatterers: now we will assume that the backscatter will be consistent at the same location at the same time of year. Using data collected at various locations in Alaska during the spring and winter, SciFish demonstrated the ability to estimate the temperature profile with less than .5 C mean error in any depth cell. Using the results, a ship-mounted sonar system was designed for use in the fisheries. Unfortunately, the Phase I funding was supplied by the Technology Reinvestment Project (TRP), and future funding for the entire TRP program was canceled, leaving no Phase II funding opportunity. As such, SciFish proposes to build upon the results of this earlier work, with an emphasis on modeling and improved data collection and a re-alignment of the design toward Navy missions. SciFish will expand upon its initial automated broadband sonar temperature estimation approach by developing an acoustic model for comparative evaluations, modifying the existing broadband sonar system and collecting better data for performance evaluations, and developing a Navy-specific design that can be built and tested during Phase II.