Low-cost scalable passive acoustic autonomous vehicle with AI

Passive acoustic recording on underwater gliders has shown promise to map fisheries, locate marine mammals, and detect vessels in protected areas. Gliders cannot operate in shallow water, necessitating offshore deployment/retrieval and restricting mapping to deeper waters; their initial purchase price and operational costs limit the number of simultaneous deployments, and post-retrieval audio file processing requirements add expense and slow scientific discovery. We propose developing Pogo, a low-cost underwater vehicle for passive acoustic recording which can be deployed and recovered from near shore, allowing collection of high-quality audio recordings in shallow waters via inexpensive and simultaneous deployments, and removing the need for ongoing operational support. A hybrid profiler with vertical and horizontal transit capability and GPS and inertial sensor navigation, Pogo will transit to its targeted location, sink and passively station on the bottom to record audio, then surface on a programmed schedule to transmit summary data and location via satellite. Pogo will use artificial intelligence and a small, customizable neural net to automatically detect signals of interest, including fish, marine mammals, or vessels, thus eliminating the need for post-retrieval audio file processing, while a cloud system and data dashboard will allow users to visualize data and manage simultaneous Pogo deployments.