Probability Modeling of Multi-Type Autonomous Unmanned Combat Aerial Vehicles Engaging Non-Homogeneous Targets Under Imperfect Information

摘要

UCAVs are advanced weapon systems that can loiter autonomously in a pack over a target area, detect and acquire the targets, and then attack them. Modeling these capabilities in a specific hostile operational setting is necessary for addressing weapons design and operational issues. While much attention has been given to the engineering and technological aspects of UCAV developments, there are very few studies on operational concepts for these weapon systems and their effectiveness and efficiency. This thesis builds probability models (Markov Chains) that describe UCAV operations, defines Measures of Effectiveness (MOEs) for the engagement performance, maps the functional relations between the parameters and the MOEs, and obtains insights regarding the design of the UCAVs and their tactical employment. The models are used to conduct extensive numerical analysis, based on experimental design concepts and traditional sensitivity analysis. The main focus of the analysis is to investigate optimal and robust mixes of UCAVs of different types, with respect to the MOEs. While in most cases, extreme-point solutions are optimal, there are cases where a balanced UCAV mix is better.