ARIES navigation system accuracy and track following

Abstract

One of the greatest challenges associated with the Autonomous Underwater Vehicle (AUV) is reliability, accuracy, and the high precision navigation system for its submerged operations. Data collected for later analysis can be meaningful if, and only if, the precise location of the vehicle is known at the time the information is recorded. A reliable AUV must be able to determine its global position in the absence of external transmitting devices. Dead reckoning is unreliable because of current conditions and random errors in the velocity measurement that can be integrated and propagated in position calculations for long distance submerged travel. The alternative is the optimal integration of all available organic vehicle sensors to determine vehicle position. This requires the Kalman filtering method which merges all available vehicle sensors to estimate position. The AUV ARIES was operated in the Azores from August 10-12, 2001. All information were recorded and transferred into several files for all the mission runs during the exercise. This thesis investigated the accuracy of the Kalman filter navigation system during those runs. The thesis also examines the actual vehicle tracks during the experiment with both the design tracks and the model prediction tracks built using a simulation of the vehicle track following behavior.