Adaptive AUV Control for Optimized Swathe Laser Stripe Imaging

Abstract

As part of a three year collaborative research project funded by EPSRC involving both Cranfield Universe and the University of Plymouth, a Deep Mobile Target (DMT5) torpedo-shaped underwater vehicle has be developed into an AUV called Hammerhead. In addition to several low cost navigation sensors, the AUV has be installed with a laser-assisted imaging and navigation sensor. This sensor is based on the laser stripe illumination (LSI) methodology previously developed at Cranfield University (Tetlow and Allwood, 1995), providing enhance viewing of the seabed below the vehicle. However, it also provides real time altitude and seabed-relative translation velocity data to the navigation system during the mission (Dalgleish et al., 2003), as well as gather images to product a post mission enhanced optical waterfall image of a surveyed area (Dalgleish et al., 2004).For a particular water type and clarity, the optimal stand-off distance (altitude and hence swathe) between the LSI a the seabed will vary. The optimisation problem is essentially the pursuit of a balance between image serve efficiency and image quality. This paper describes the development of an automatic process for monitoring the I image quality to determine the optimum altitude at which to fly, and henceforth instruct the AUV control system do so. In this way, swathe may be optimised for maximum seabed coverage and more efficient use of AUV time. 1 paper also presents simulation results from a genetic algorithm (GA) based model predictive control (MI methodology used to track the required altitude provided by this process.