Graphical Processing Unit (GPU) Software to Accelerate Underwater Acoustic Autonomous Modeling and Processing

3 Phoenix, Inc. proposes to develop and evaluate technology that provides a substantial improvement in processing capability for autonomous sonar processing. Improved signal processing, automated classification and propagation models place additional loads on battery power. 3Pi proposes to demonstrate feasibility of a relevant software application running in hardware using 7.5-10 watts. We propose to compute hardware benchmarks in order to evaluate processing/power trade-offs. The proposed approach leverages commercial-off-the-shelf (COTS) processor hardware. We have significant experience developing low power implementations of signal processing and automated classification software and firmware, and the proposed approach for the Phase I research is to leverage previous work. We anticipate that experience at 3Pi in this field will allow successful development of solutions appropriate for multiple Navy problems.