Feature Based Localization and Navigation for Miniature Underwater Vehicles

Significant resources are required to ensure proper self-localization of submersibles without available reference signals such as GPS or LBL. This is compounded, when the vehicle remains submerged for extended periods of time which is often required for reconnaissance missions. In the case of a miniature autonomous underwater vehicle (AUV), the self-localization challenge extends well beyond INS filtering. Although, position and velocity sensor systems (DVLs) have been successfully designed for large and medium scale submarines, the miniaturization of those sensor components still poses a significant challenge. To close this technological gap, the NextGen team proposes the development of a sonar-based miniature navigation sensor system that aims to challenge the weight, size, cost as well as performance specifications imposed by small and miniature AUVs. The proposed solution will combine advanced sonar technology with software modules capable of deducing the vehicle kinematics, i.e. position and velocity, in real time. The NextGen, Virginia Tech and Lockheed Martin team combines year-long expertise in i) sensor calibration and underwater navigation, ii) system integration, and iii) image processing and state estimation. For performance demonstration, the team plans to integrate the sensor into either VTs 690 AUV or NextGens Unmanned Underwater Riverine Craft (UURC).