Using Discrete Event Simulation to Assess Obstacle Location Accuracy in the REMUS Unmanned Underwater Vehicle

摘要

Navy personnel use the REMUS unmanned underwater vehicle to search for submerged objects. Navigation inaccuracies lead to errors in predicting the location of objects and thus result in increased search times for Explosive Ordnance Disposal (EOD) teams searching for the object post-mission. This thesis explores contributions to navigation inaccuracy using Discrete Event Simulation (DES) to model the vehicle's navigation system and operational performance. The DES produced for this thesis uses the JAVA-based Simkit package to simulate the navigation system in REMUS. The model considers factors affecting accuracy, such as compass error, the effect of current, transducer drop error, transducer positioning effects, and ping interval. Mines can be placed at specific locations or generated randomly. Three types of vehicles are considered in this thesis. First, a simple vehicle that navigates by Dead Reckoning is analyzed. Second, a more complex vehicle that navigates using Long-Baseline (LBL) is analyzed. Third, the vehicle is simulated to move through an area of interest in a sweeping pattern that is populated by 10 mines, each of which is randomly positioned. Data from the last vehicle are used to build three analytic models that the operator can use to improve performance. First, the probability of detection is modeled by a logit regression. Second, given that detection has occurred, the mean location offset is modeled by a linear regression. Third, the distribution of errors is shown to follow an exponential distribution. These three models enable operators to explore the impact of various inputs prior to programming the vehicle, thus allowing them to choose the best combination of vehicle parameters that minimize the offset error between the reported and actual locations. (18 tables, 33 figures, 17 refs.).