| Summary: | Submitted in partial fulfillment of the requirements for the degree of Master of Science in Mechanical Engineering at the Massachusetts Institute of Technology and the Woods Hole Oceanographic Institution September 2021. A novel performance metric to improve underwater digital acoustic communication, called Multipath Penalty (MPP), is proposed as an alternative to traditional signal-to-noise ratio (SNR) methods in the context of the Arctic Beaufort Sea. MPP and SNR are compared alongside a third performance metric, Minimum Achievable Error (MAE), which replicates the operation of a channel estimate-based decision feedback equalizer in an acoustic modem. The three metrics are then tested in a hardware-in-the-loop Virtual Ocean simulator for an autonomous undersea vehicle (AUV) communicating with a collaborator. Using field data of modem statistics obtained duringICEX20 and expanded data supplied by the simulator, calibration of the three metrics to modem packet success is evaluated, resulting in a proposed recalibration for MAE. The AUV’s ability to communicate when adaptively choosing its depth is analyzed above and below the Beaufort Lens, and settings for MPP’s engineering variables are obtained. The results show MPP generally improves reception and demodulation of acoustic transmissions over SNR by approximately 5% within an operational range of 8 km, while achieving similar results to the more robust metric MAE. MPP is an improved utility for underwater digital acoustic communication in both marine autonomy and as a tactical decision aid. This work would not be possible without the extraordinary support of the United States Navy, which provided funding for this research, my degree, and my livelihood as an active duty submarine officer.
|
|---|