Applying Utility Theory to Improve Autonomous Underwater Vehicle Planning Around Failed Radio Frequency Antennae

摘要

Militaries, commercial organizations, and research organizations are increasingly relying on unmanned vehicles to perform missions that are dull, dirty, dangerous, or otherwise cost prohibitive. Whereas surface and air vehicles can readily communicate with external information sources, such as human operators or satellites, undersea vehicles have limited communication capabilities, due to the electromagnetic signal-attenuating properties of water. As a result, autonomous underwater vehicles (AUVs) must be capable of robust decision-making. To minimize the occurrence of aborted missions, this paper presents a method by which utility theory can be applied to the decision-making processes of the vehicles for mission planning and replanning. Specifically, this paper focuses on mission replanning around a failed communications antenna. Such an application allows vehicles on such a mission to determine the 'best' alternate payload to use without operator intervention, thus improving vehicle reliability. Because 'best' is subjective, focusing on relevant payload attributes and tailoring these functions to individual operator preferences ensures a unique vehicle makes decisions that align with a unique operator's preferences. The utility functions presented in this paper can be expanded to take into consideration attributes in addition to those already addressed. Further, utility functions for additional missions and additional types of vehicles can be created, using the process outlined in this paper. Once this new set of expanded utility functions is created, the functions can be inserted into the mission planning and replanning logic of an AUV, allowing that AUV to more autonomously make decisions without the assistance of a human operator.