In the last decade, vision systems have become an
instrumental sensor in the development of accurate control and navigation
systems for unmanned underwater vehicles. Unfortunately testing of new vision
based techniques in actual underwater environments is a difficult and expensive
proposition. This paper details the development and evaluates the potential
performance of two novel methods for underwater vehicle vision based control
and navigation systems testing; physical simulation and virtual simulation. The
paper also details the experimental setup for test tank trials utilizing a
seabed poster representation. This setup is evaluated based on results from a
previously developed real-time vision based navigation system. The experimental
setup for sea trials of the University of Limericks newly developed thrusted
pontoon/ROV is also presented.