Terrain-Based Localization and Mapping for Autonomous Underwater Vehicles using Particle Filters with Marine Gravity Anomalies

The feasibility of using prior gravity anomaly measurements that are 1 nautical mile apart for underwater simultaneous localization and mapping is studied. This paper reports on modelling and simulation that investigates relationships between the characteristic anomaly parameters and the requirements for particle filter SLAM solutions. Map anomaly variability parameters suggest the gross likelihood of SLAM success in a map. However, the anomaly parameters that relate to the local anomaly measurement-to-measurement variability (i.e. localization) are better indicators for a SLAM mission's success. The prior gravity anomaly measurements, coupled with the tools developed here, provide guidance to select optimal areas and missions the AUV could transit through towards minimal localization error at the goal location. Follow-on work will exploit the anomaly localization parameter to assist with SLAM path-planning through a vector field histogram approach and then, implement it on an AUV for validation. (C) 2018, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.