The application of multi sensor detectors and/or sophisticated signal processing systems are two of the most important developments in fire detection technology. Due to the increasing complexity involved the layout of fire detection algorithms becomes more and more elaborate. Based on a representative set of recorded sensor signals layout procedures and performance studies of fire detection algorithms can be aided by simulation techniques. This requires adequate and efficient modelling of the signal recordings by a stochastic model. This paper presents a modelling technique based on a vector autoregressive (VAR) model. This model takes into account the cross correlations between the single sensor signals and is therefore well applicable in the field of automatic fire detection. It can serve as a basic component of a development tool for the design of detection algorithms in automatic fire detection technology. The applicability of the model will be shown by simulation results for signals of a ionization chamber, an optical scattered light smoke sensor and temperature sensor used for automatic fire detection measurements in the field.
[1]
G. C. Tiao,et al.
An introduction to multiple time series analysis.
,
1993,
Medical care.
[2]
William W. S. Wei,et al.
Time series analysis - univariate and multivariate methods
,
1989
.
[3]
Richard A. Davis,et al.
Time Series: Theory and Methods
,
2013
.
[4]
Peter Strobach,et al.
Linear Prediction Theory: A Mathematical Basis for Adaptive Systems
,
1990
.
[5]
Ulrich Appel,et al.
Adaptive sequential segmentation of piecewise stationary time series
,
1983,
Inf. Sci..
[6]
Sune Karlsson,et al.
Introduction to multiple time series : H. Lutkepohl, 1991, (Springer, New York), 552 pp., paperback US$59.00, ISBN 0-387-53194-7
,
1993
.
[7]
Helmut Lütkepohl,et al.
Introduction to multiple time series analysis
,
1991
.