Repository hosted by TU Delft Library

Home · Contact · About · Disclaimer ·
 

Extending the COVAD toolbox to accommodate system non-linearities

Publication files not online:

Author: Bucco, D. · Weiss, M.
Type:article
Date:2009
Institution: TNO Defensie en Veiligheid
Source:AIAA Modeling and Simulation Technologies Conference, Chicago, IL, USA, 10-13 August 2009 (Conference code: 81962)
Identifier: 425198
ISBN: 9781563479786
Article number: No.: 2009-6018
Keywords: Weapon systems · Adjoints · Analytical techniques · Covariance analysis · Dynamic Systems · Guided missiles · Linearization technique · MATLAB /simulink · Miss distance · MONTE CARLO · Nonlinear problems · Nonlinearities · Rapid analysis · Statistical linearization · Computer simulation · Linear systems · Linearization · Monte Carlo methods

Abstract

The COVAD toolbox is a MATLAB/Simulink based tool conceived and developed for the rapid analysis and simulation of stochastically driven dynamic systems. In addition to a generic Monte Carlo capability, the toolbox is also supported by traditional analytical techniques such as the adjoint and covariance analysis methods. However, these latter techniques only apply to linear systems. The objective of this paper is to explore the feasibility of extending COVAD to enhance its non-Monte Carlo capability for the analysis of non-linear systems. The present investigation will focus on the use and implementation of a well-known linearization technique known as the statistical linearization method. This technique has been used in the past in conjunction with the standard adjoint and covariance analysis method to provide sufficiently accurate solutions to certain classes of non-linear problems. Calculating the miss distance statistics of the homing loop of a generic guided missile under acceleration limiting will then be used as an example to demonstrate the utility of the software. Copyright © 2009 by the American Institute of Aeronautics and Astronautics, Inc.