Erik Andersson och David Börjesson
	
Statistical Validation of IR Signature Software

Abstract

This report describes a statistical validation of a software used
by the Swedish Defence Research Agency to simulate IR signatures
of specific objects. The differences found between simulated
results and real measurements can often be attributed either to
model deficiencies or to uncertainties in the input parameters
used by the software. We associate suitable uncertainties to each
input and investigate how this affects the output, using three
different methods. In addition to this, we determine which
parameters contribute the most to the output variance.