Analysis and Tuning of a Low Cost Inertial Navigation System in the ARIES AUV

Autonomous underwater vehicle navigation is a complex problem of state estimation. Accurate navigation is made difficult due to a lack of reference navigation aids or use of the Global Positioning System (GPS) that could establish the vehicles position. Accurate navigation is critical due to the level of autonomy and range of missions and environments into which an underwater vehicle may be deployed. Navigational accuracy depends not only on the initialization and drift errors of the low cost Inertial Motion Unit (IMU) gyros and the speed over ground sensor, but also on the performance of the sensor fusion filter used. This thesis will present the method by which an Extended Kalman Filter (EKF) was tuned after installation of an IMU in the ARIES Autonomous Underwater Vehicle. The goal of installing the IMU, analyzing the navigational results and tuning the EKF was to achieve navigational accuracy in the horizontal plane with a position error of less than one percent of distance traveled when compared with GPS. The research consisted of IMU installation and software modifications within the vehicle to fully realize the design goal. Data collection and analysis was conducted through field experiments and computer simulation. A significant result of this work was development of a pseudo-adaptive algorithm to vary the measurement noise values in selected channels to force a desired response in the filter and improve accuracy and precision in the state estimates.