Decentralized Single-beacon Acoustic Navigation: Combined Communication and Navigation for Underwater Vehicles

摘要

This thesis reports the derivation and validation of two single- beacon acoustic navigation algorithms, as well as the development and experimental evaluation of a platformindependent acoustic communication (Acomms) system that enables combined communication and navigation. The navigation algorithms are centralized and decentralized formulations of single- beacon navigation, and employ range measurements from a single reference beacon to an underwater vehicle in addition to the vehicles Doppler velocity log gyrocompass, and depth sensors to perform absolute (as opposed to relative) localization and navigation of the vehicle. The centralized single-beacon algorithm is based on the extended Kalman filter. We assume that the Kalman filter has simultaneous, real-time access to sensor measurements from both the vehicle and the beacon (e.g. the ship). The decentralized single-beacon algorithm is based on the information form of the extended Kalman filter. We assume that the information filter on the vehicle only has access to measurements from the vehicles on-board navigation sensors in real-time. The vehicle-based filter receives acoustic broadcasts from the reference beacon that contain information about the beacons position and sensor measurements. We show analytically and in simulation that the decentralized algorithm formulated herein yields an identical state estimate to the state estimate of the centralized algorithm at the instant of each range measurement; in addition we show that between range measurements the results from the two algorithms differ only by linearization errors and the effects of smoothing historic ship states. The Acomms system has been installed on theWood Hole Oceanographic Institution vehicles Puma, Jaguar and Nereus. The author and collaborators deployed the Acomms system in four sea trials in the North Pacific and South Atlantic Oceans.