Improving AUV Localization Accuracy by Combining Ultra-Short-Baseline and Long-Baseline Measurements Systems in a Post-Processing Extended Kalman Filter

Abstract

This paper presents and evaluates an approach for combining ultra-short-baseline (USBL) and long-baseline localization data in a post-processing extended Kalman filter. The goal is to improve localization accuracy and reduce localization uncertainty for autonomous underwater vehicles (AUVs) performing oceanographic survey measurements. The method was evaluated using logged LBL navigation data from field testing and simulated USBL data. Localization accuracy was evaluated by comparing state uncertainties of the independent and combined systems. Uncertainties of USBL localization data were estimated using a Monte-Carlo simulation at each AUV position. The results indicate that combining USBL data improves localization accuracy, especially when the USBL data includes depth telemetry measurements. Although this approach was evaluated by adding USBL data to logged LBL field-testing data, it could be applied to any navigation approach, including dead-reckoning.