Date of Award
Doctor of Philosophy (PhD)
Department of Operational Sciences
Kenneth W. Bauer, Jr., PhD
This dissertation research makes contributions towards the evaluation of developing Automatic Target Recognition (ATR) technologies through the application of decision analysis (DA) techniques. ATR technology development decisions should rely not only on the measures of performance (MOPs) associated with a given ATR classification system (CS), but also on the expected measures of effectiveness (MOEs). The purpose of this research is to improve the decision-makers in the ATR Technology development. A decision analysis framework that allows decision-makers in the ATR community to synthesize the performance measures, costs, and characteristics of each ATR system with the preferences and values of both the evaluators and the warfighters is developed. The inclusion of the warfighter's perspective is important in that it has been proven that basing ATR CS comparisons solely upon performance characteristics does not ensure superior operational effectiveness. The methodology also captures the relationship between MOPs and MOEs via a combat model. An example scenario demonstrates how ATR CSs may be compared. Sensitivity analysis is performed to demonstrate the robustness of the MOP to value score and MOP to MOE translations. A multinomial section procedure is introduced to account for the random nature of the MOP estimates.
DTIC Accession Number
Bassham, Christopher Brian, "Automatic Target Recognition Classification System Evaluation Methodology" (2002). Theses and Dissertations. 4361.