Abstract
An important application for Autonomous underwater vehicles (AUVs) is to utilize them to survey large underwater areas to detect objects of interest. The capabilities of AUVs are augmented with the multi-objective autonomy platform called the Mission Oriented Operate Suite (MOOS), Interval Programming (IvP), which enables behavior-based autonomy for robotic platforms. An Expandable Decision Framework was created that supports real time post processing of side-scan sonar data, feature detection within the data, and facilitates action regarding detected features. The framework has been implemented to utilize external modules for the post processing of raw sonar data, and feature detection of the processed sonar data, such that different implementations can be used for different techniques. Modules for the specific use case of performing dynamically created re-acquire and identify (RID) search patterns on localized anomalies that are detected during a survey mission have been developed to prove the framework is capable of real-time analysis and action. The framework has been tested in simulation with multiple previously collected sonar data to simulate AUV missions.