We present an approach to convert slow running, high fidelity simulations into fast running models via analogy. These high order polynomial models use Data Modeling coupled with multivariate regression. Shown are models for individual data sets, field measured data, simulations like the Navy Experimental Dive Unit rebreather model, simulation driver equations for the Kinetic Impact Debris Distribution model, and a model of a terahertz sensor sensing Bt simulant built from a hardcopy only. These equation-based models are easily embedded into larger system simulations as active model components for enabling system performance and trade analysis. Models are fast running, and in the case of simulation and driver equation modeling, provide results that are significantly shorter in lines of software code.

Date Published: 25 October 2002
PDF: 12 pages
Proc. SPIE 4772, Electro-Optical System Design, Simulation, Testing, and Training, (25 October 2002);

Published in SPIE Proceedings Vol. 4772:
Electro-Optical System Design, Simulation, Testing, and Training
Richard M. Wasserman; Scott L. DeVore, Editor(s)