COOPERATIVE LOCALIZATION AND BATHYMETRY-AIDED NAVIGATION OF AUTONOMOUS MARINE SYSTEMS

Abstract

Localization and navigation are the key components determining the performance of underwater autonomous vehicles (AUVs). Traditional localization either equips a single AUV with high-end navigation systems, or assists AUV with installation of external beacons. This thesis focuses on a team of small-sized and low-cost AUVs, where no single AUV functions as a beacon possessing accurate position. The team is able to outperform single-vehicle operation through cooperation. Due to the underwater communication constraints, it is impractical to collate all sensor data centrally, and therefore a decentralized processing architecture is preferred. We design a feasible distributed localization algorithm, which records the correlated terms and requires minimum communications. As naïve assumption is commonly used, we show how detrimental it is and quantify the safe region. With this quantification, we proceed to path planning with bathymetric aids for a single AUV.