Link Foundation Ocean Engineering and Instrumentation Fellowship Reports
Recent advances in unmanned underwater vehicle (UUV) capabilities have enabled oceanographic scientists to consider addressing research topics previously thought impractical or impossible. One such topic is the validation of high resolution computer models; long term deployments of coordinated UUV survey teams have been proposed to collect the density of measurements required. Another topic is the study of biological diversity in volatile areas; UUV intervention/sampling deployments have been proposed because of the inherent risks. In general, providing a high density of high accuracy measurements requires passing many targets as quickly and closely as possible. Similarly, efficacy for intervention/sampling deployments requires agile and accurate vehicle operation. Thus full utilization of UUV capabilities for ocean sensing, and therefore ocean science, requires accurate trajectory tracking. Currently UUVs use linear controllers which do not model or compensate for nonlinearities such as drag, buoyancy, inertia and hydrodynamic forces. In experimental evaluations of UUV Model Based Controllers (MBC), compensating for such nonlinearities provided significant performance gains over linear controllers . Since the system identification necessary for MBC is rarely available, I feel Model Based Adaptive Controllers (MBAC) will be required to realize these performance gains because MBACs are able to learn the correct model parameters as part of the control process.
McFarland, Christopher J., "Model Based Adaptive Control of Underwater Vehicles for Improved Ocean Science" (2013). Link Foundation Ocean Engineering and Instrumentation Fellowship Reports. 19.