EAGER: Toward Descriptive Mapping for Underwater Exploration

This project will provide a new solution for 3D mapping and exploration of underwater environments characterized by sparse and noisy data. It will enhance the capability of underwater robots to autonomously map, navigate, and inspect a previously unknown shallow-water environment. The development of this project will benefit from its long association with United States Coast Guard. This project will advance the state of the art in three-dimensional occupancy mapping using Gaussian process regression, a supervised learning method with great promise for its application to robot mapping. This project will apply these methods to online 3D mapping with a field robot, mapping underwater structures with a scanning sonar.