In order to do anything about environmental problems, we must first understand the environment. We cannot understand the environment without a thorough understanding of the ocean, yet it is very difficult to study. Autonomous underwater vehicles, especially long-range AUVs, offer the best hope for gathering the kind and amount of data about the ocean needed for environmental understanding. However, AUVs require intelligent controllers if they are to carry out the kind of complex, possibly long-duration missions necessary.
The Orca project is taking a phased approach to creating an intelligent AUV controller. The first version of Orca is nearing completion. It will, as mentioned above, make use of rule-based systems for some components (though not for planning). It remains to be seen during evaluation if we will need to replace the rule-based reasoners with other techniques. Future versions of Orca will incorporate increased knowledge and reasoning abilities, including enhanced temporal and spatial reasoning abilities.
The first few versions of Orca will be developed and tested in the simulation testbed [\protect\citenameTurner et al., 1991]. Later versions will be fielded aboard the MSEL EAVE and long-range AUVs for in-water tests. Ultimately, the program will be used to control the AUVs during ocean science missions.