Abstract

The growing use of autonomous air, surface, ground and underwater systems is continually demonstrating new military and commercial possibilities and applications. The state of the art in the planning and control of autonomous underwater vehicles (AUVs) is largely precarious because AUVs provide infrequent feedback, operate autonomously for long periods of time and yet have little knowledge of their dynamic environment. Consequently, mission planning and control is typically conducted based on human expert knowledge of vehicle capabilities, some level of observed environmental conditions and ad-hoc optimization with little assistance from computers. While the human expert offers a significant ability to mentally process data, the result typically lacks numeric and quantitative analysis of alternatives. Navigation of AUVs in the complex ocean environment involves time dependent dynamics, resulting in a problem that is computationally prohibitive for the use of brute force optimization techniques. Although some research has been conducted for specific types of missions, and "greedy" global-optimization approaches have been investigated, no systematic and coherent approach to the requirement exists. We propose a dynamic multiple criteria decision support system (DSS) that considers dynamic and episodic ocean phenomenon to provide reasonable and in-context recommendations with respect to the stated objective and subjective mission goals. Multi-criteria decision analysis (MCDA), analytic network process (ANP) and fuzzy sets are used in the model to reduce the vehicle routing solution space and maximize time-on station in adverse environments. The proposed system can be an added hierarchical layer on the top of a mission planning system currently under development by the United States Navy.