Towards a Multi-Modal Navigation Scheme for Deep Submergence Robotic Vehicles

Modern fully aided Inertial Navigation Systems (INS) enable precise navigational capabilities and entirely new concepts of operation for underwater robotic platforms. These INS systems utilize inputs from Ultra Short Baseline (USBL), Doppler Velocity Log (DVL) bottom and water track, GPS position, sound velocity, and depth information to dramatically improve the real-time navigation accuracy of the vehicle. However, many INS systems suffer from the “black box” problem wherein the lack of transparency in the process of fusing many inputs to form an interconnected navigation “solution” can lead to propagating errors and other issues in operational environments. As such, the development and qualification of a new INS-based navigational system across robotic platforms must consider individual platform sensor accuracy, reliability, and operator usability. The new multi-layered navigation system developed and integrated on National Deep Submergence Facility (NDSF) vehicles utilizes a sensor-aided INS while supporting legacy dead reckoning navigation, secondary Attitude and Heading Reference System (AHRS) sensors, and providing transparency in inputs, outputs, and real-time performance.