• Autonomous Underwater Vehicles
• AUV
• extended kalman filter
• GIB system
• inverted LBL
• maximum likelihood estimator
• outlier rejection
• trilateration
• underwater positioning systems

This thesis proposes several solutions to the problem of estimating the position of an underwater
vehicle. The commercial system adopted for collecting data consists of four buoys
that compute the times of arrival (TOA) of the acoustic signals emitted periodically by a
pinger installed on-board the target (so called G.I.B. system).
The solutions discussed are grouped in two family of algorithms: instantaneous positioning
(static) algorithms and filtering (dynamic) algorithms. For the static group, the
Trilateration problem is discussed in depth, yielding closed form and iterative solutions. A
detailed discussion of the estimation accuracy of the algorithms is presented.
For the dynamic group, an EKF based on a simple kinematic model of the target was
introduced. In order to test the algorithms proposed with real data from sea trials an
initialization procedure was implemented. The procedure provides a initial fix that lies
inside a certain confidence region. The filtering algorithms can be dramatically affected by
outliers. Thus, an on-line measurement validation was also introduced.
A Matlab-based graphical user interface was designed in order to test the algorithms
described with real data obtained during sea experiments.