Improving Position-Based Alignment for AUV's Localization Through Optimal Data Selection

This paper describes a method for improving the performance of the position-based in-motion alignment algorithm used to accurately estimate the initial heading of AUVs(Autonomous Underwater Vehicles). The alignment algorithm in prior research used GPS(Global Positioning System) positions and navigation-estimated positions to correct the heading angle. However, significant errors in some GPS data adversely affected the accuracy of heading bias estimation. To resolve this issue, this study divides the position data into candidate data and validation data. The optimal sample for estimating heading bias is then identified from the candidate data using the validation data. The proposed algorithm is verified through experiments using sensor datasets acquired from real sea experiments.