Automated Ice Detection for Autonomous Underwater Gliders

Autonomous underwater vehicles (AUVs) and gliders (AUGs) face significant operational challenges in ice-covered regions due to the need for periodic surfacing to transmit data, update position estimates, and facilitate recovery. In these environments, vehicles can become trapped, crushed, or displaced by currents or drifting ice. While large-scale distribution of sea ice and available open water can be detected via satellite imaging, it is not practical to communicate this information in a timely manner to underwater vehicles while submerged. Smaller, meter-scale ice fragments can pose significant risks to vehicle survivability but are beyond the spatial and temporal resolution capabilities of earth observing satellites. In this work, we describe an automated ice detection method using data collected by an AUG equipped with an upward-facing Mechanically Scanned Imaging Sonar (MSIS). We demonstrate the efficacy of this approach with results from field data collected in the Arctic marginal ice zone (MIZ).