|
Title:
|
EFFICIENT WORKING AREA EXPLORATION FOR DATABASED MODELING OF AUTOMOTIVE APPLICATIONS |
|
Author(s):
|
Philipp Senger , Holger Mielenz , Rolando Doelling , Wolfgang Rosenstiel |
|
ISBN:
|
978-972-8924-62-1 |
|
Editors:
|
Hans Weghorn and Ajith P. Abraham |
|
Year:
|
2008 |
|
Edition:
|
Single |
|
Keywords:
|
Black Box behavior modeling, A/MS systems, Automated and adaptive working area exploration, Latin Hypercube
Sampling, Analysis of parameter sensitivity, Au |
|
Type:
|
Full Paper |
|
First Page:
|
67 |
|
Last Page:
|
74 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
This paper presents an extension of a black box behavior modeling by an automated working area exploration approach
for analog and mixed signal (A/MS) systems. The behavior modeling is based on the transient data tuple of input and
output data. The used statistical modeling strategy are the Support Vector Machines (SVM). The so far manual data
generation and the choice of the relevant data for the modeling should be automated by two methods. The first attempt is
an a priori method which defines the parameter of the relevant input signals for the stimulation of the system by a latin
hypercube sampling (LHS) logic. The LHS approach is extended afterwards by an active selection of the stimulus
parameters over a parameter sensitivity analysis (PSA). This method identifies the sensitive ranges of a system
autonomously and defines further simulation experiments, in order to generate aimed data. Whereupon a uniform
stimulation environment is pointed out which affords the automation of these methods. Finally the introduced methods
are tested and discussed by an automotive example. |
|
|
|
|
|
|
