Babel Shark: Improved Bridge-to-Bridge Radio for Unmanned Surface Vehicles

R-DEX Systems (R-DEX) proposes to team with the Georgia Tech Research Institute (GTRI) to develop Babel Shark, an innovative system that converts VHF Bridge-to-Bridge radio transmissions from voice to text to meaning and integrates them into a COLREGS reasoning engine, thereby generating an intelligent reply to the proposed maneuver. In Phase I, the R-DEX/GTRI team will address the most significant challenges in transitioning speech-based communication in USVs into the field, namely: 1) Integration of a speech-text-meaning-speech engine with mature, fielded software; 2) enhancement of robustness through algorithms for English-as-a-second-language (ESL) speakers and machine-learning-enabled speech signal enhancement; 3) rigorous performance quantification. To address these challenges, Babel Shark will require several innovations. The maximum technological impact will hinge on developing low-SWAP instantiations so that it is compatible with the greatest number of platforms. The incorporation of advanced speech signal enhancement techniques and an ESL model will improve the robustness, further improving the chances of technology insertion into a high-consequence application. At the end of Phase I, the team will have developed an end-to-end prototype, integrate the solution into a fielded software (FalconView) that serves as a COLREGs surrogate, and demonstrate the ability to perform speech enhancement in a relevant noise environment.