| SnoMotes: Arctic Robots | Under-Ice Robotics | Underwater Human-Robot Interaction |
Many fields of science (including biology, climate, and planetary) are interested in exploring under the polar ice. Harsh under-ice environments though restrict access to divers and manned submersibles. Thus, the use of unmanned under-ice vehicles (UUVs) are ideally situated for scientific data collection. Current common underwater navigation solutions though have limitations under the ice – UUVs can’t surface thus there is no availability of GPS; acoustic beacon methods require prohibitive setup and infrastructure; UUVs also suffer from inertial drifts and compass singularities; and the environment itself is fairly featureless. As such, in this research, we focus on developing computer vision methods with acoustic and optical imaging sensors, to enhance under-ice UUV motion estimation, as well as automatic ice texture and ice anomaly mapping, despite the mostly featureless environments under the ice. This project was completed in conjunction with Dr. Mick West (GTRI) and Dr. Britney Schmidt (GT Earth & Atmospheric Science)
