Analysis of the Effects of Dynamic Characteristic Dimension Calculations on FLIRPerformance Prediction Models

摘要

The ability to accurately predict the performance of FLIR systems has becomecritical to today's military. The current U.S. defense industry standard FLIR analysis model is FLIR92 by the U.S. Army's C2NVEO. The algorithm in FLIR92 for calculating target acquisition probabilities, called ACQUIRE, has several limitations in its design for calculating a target's characteristic dimension. This thesis develops a Dynamic Model to overcome these limitations. It incorporates a three dimensional view of a target based on range, azimuth angle to target, and the altitude of the FLIR sensor. An analysis of the effects of dynamically calculating a target's characteristic dimension by the Dynamic Model and the static ACQUIRE version 1 model is presented. Both are compared on a theoretical target from three different angles; the front, the 45, and the side, with the Dynamic Model producing an 8% increase in prediction ranges for the front, a 4% increase for the 450 view, and a 5% increase for the side. An Empirical Cumulative Tail Distribution is computed from experimental data, and the theoretical probability vs. range predictions of each model are then compared to actual observations. A sensitivity analysis is performed to demonstrate the effects of various conditions on predicted acquisition ranges.