Experimental Setups for Vision-Based Navigation of Unmanned Underwater Vehicles in Near Intervention Operations

Abstract

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.