| Summary: | This thesis investigates the combined use of ocean models, such as idealized surface current flows, and search models, including expanding area and discrete myopic search methods, to improve the probability of detecting a near-surface, drifting object over time. Enhanced search effectiveness is facilitated by the use of robotic search agents, such as a tactical unmanned underwater vehicle (UUV) or unmanned aerial vehicle (UAV), leveraging simulation methods to inform the search process. The presented work investigates the impact of using nave versus optimized search patterns on localizing a drifting object, including a surrogate ocean model using idealized flow as well as historical data sets with Weibull-distributed perturbations. Numerical studies and extensive analysis using different permutations of model parameters (including the relative speed of the drifting object, time late in the searcher�۪s arrival to the search area, sensor sweep width, and duration of the search mission) identify the significant factors affecting the overall probability of detection. Such insights enable further explorations using empirical datasets for specific oceanographic regions of interest.
|
|---|
