Mapping the seafloor is a basic first step for much oceanographic research. Autonomous Underwater Vehicles (AUVs) equipped with multibeam sonar systems have launched a revolution in ocean science and exploration because these platforms a both stable and efficient; they have been used for a host of applications from mapping sand ripples and seamounts to sea ice. However, producing accurate large scale maps from the AUV data is still a highly expert labor intensive activity. The objective of this project is to develop computer program to enable rapis and efficient production of bathymetric maps from AUV data. The chief broader impact of this work is that it will provide dramatically improved tools for mapping the seafloor using AUVs for the entire oceanographic community.
This project focused on created accurate and self consistent bathymetry maps using underwater robotic vehicles. Such maps are produced by processing multibeam sonar data and vehicle navigation data. These data are collected when the vehicle flies a survey pattern at altitudes ranging from 3 to 50 meters off of the bottom. When producing maps it is challenging to repeatably and consistently align the data in a manner free of artifacts related to poor positional navigation estimates. Such artifacts manifest themselves as a lack of clarity in the map, and sharp or distinct patterning that is not related to the underlying sea floor. This project has furthered the development of an automated approach to produce better maps without requiring user supervision. There are many steps in automating map processing. This work focused primarily on the formulation of an automated terrain matching technique that can detect locations that have been mapped previously and use the overlap to correct for navigation errors. The end result is a better map with fewer artifacts and errors. The main contribution of this work has been reformulating the optimization process in a manner that runs much quicker and has no immediate limitations related to producing large maps. As an example, the newly developed algorithm can reproduce a particular sample map in 2.5 minutes that previously took 15 hours using an older formulation of the problem. The approach also has a structure that is conducive to merging surveys taken at different times. On going work is restructuring the developed code to work with the existing processing tool MB System. The data used for this project was collected with the Hercules remotely operated vehicle on the E/V Nautilus and with the National Deep Submergence Facility vehicle Sentry. The resulting technique is now being further refined and will become the standard map making approach for data collected using the Hercules vehicle. We are also reprocessing Sentry data to further evaluate the method. The method has also been used to produce very high resolution maps (centimeter gridded) with a bathymetric structured light laser system developed by the PI. The approach is fairly general and will work with both laser and sonar data. The publications listed below contain maps produced using the algorithm developed with this award. This project also funded a masters thesis. Kelly, J., Carey, S., Croff-Bell, K., Rosi, M., Pistolesi, M., Roman, C., Marani, M. "Exploration of the 1891 Foerstner submarine vent site (Pantelleria, Italy): insights into the formation of basaltic balloons". Bulletin of Volcanology, Under Review James Ian S. Vaughn, "Microbathymetry using self-contained navigation and simultaneous localization and mapping" MS Thesis 2012 Inglis, G. Smart, C., Vaughn, I., Roman, C., "A Pipeline for Structured Light Bathymetric Mapping", Proceedings of IROS 2012, October 7-12, Vilamoura, Algarve, Portugal 2012 Douillard, B., Nourani-Vatani, N., Johnson-Roberson, M., Williams, S., Roman, C., Pizarro, O., Vaughn, V., Gabrielle, I., "FFT-based Terrain Segmentation for Underwater Mapping", Proceedings of Robotics: Science and Systems, July, Sydney, Australia 2012 Roman, C., Inglis, G., Vaughn, I., Smart, C., Douillard, B., Williams, S., "The Development of High-Resolution Seafloor Mapping Techniques'', in Bell, K. L. C., Elliot, K., Martinez, C. and Fuller, S. eds. 2012 New Frontiers in Ocean Exploration: The E/V Nautilus and NOAA Ship Okeanos Explorer 2011 Field Season. Oceanography 25(1), supplement, 72pp.
