A post-processing Kalman smoother for underwater vehicle navigation.

摘要

This thesis describes an automated post-processing tool, designed for use on navigational data gathered by Autonomous Underwater Vehicles (AUVs), developed and operated by the Department of Ocean Engineering at Florida Atlantic University. The post-processing tool consists of a 9-state complementary Kalman filter in conjunction with a Rauch-Tung-Striebel (RTS) smoothing algorithm. The Kalman filter is run forward in time to merge navigational data from an Inertial Measurement Unit (IMU), a Doppler Velocity Log (DVL), a magnetic compass, a GPS/DGPS system and an Ultrashort Baseline (USBL) tracking system. Subsequently, the RTS smoothing algorithm is run backwards in time to find and compensate for drift errors in dead reckoned position and compass measurement error. The post-processing tool has been implemented as a graphical user interface, designed in MATLAB. Improved accuracy in post-processed position and heading has been verified by conducting sea trials and post-processing the collected data.