Previous Abstract Return to Session D8 Next Abstract

Session D8: AI/Machine Learning: MagNav

MINIMAN
Spencer Low, Brigham Young University; Isaac Ege, University of Dayton Research Institute; TJ Gaydosh, University of Dayton; Dylan Bowald, Air Force Research Laboratory; Enjie Wang, Cornell University
Location: Ballroom E
Date/Time: Wednesday, Jun. 5, 8:55 a.m.

This paper describes the project MINNIMAN, or Map Inferencing Neural Networks Improving Magnetic Anomaly Navigation. It takes recent work in image generation, and in particular paired image to image translation, and attempts to apply them to the task of inferring magnetic anomaly maps. We accomplish this using a novel architecture based on a modified VQGAN (Vector Quantized Generative Adversarial Network) architecture, with a notable addition of an auto-encoding pretraining step for faster translations between multiple domains. We also explore its application to inferring other alt-nav reference maps.

Previous Abstract Return to Session D8 Next Abstract