Title :
Quadratic tests for detection of abrupt changes in multivariate signals
Author :
Nikiforov, Igor V.
Author_Institution :
Dept. de Genie des Syst. d´´Inf. et de Decision, Univ. de Technol. de Troyes, France
fDate :
9/1/1999 12:00:00 AM
Abstract :
This article considers the problem of detecting abrupt changes in the mean of a multivariate Gaussian random signal. A fixed sample size χ2-test is compared against the optimum sequential tests (χ2-CUSUM and χ2-GLR)
Keywords :
Gaussian processes; maximum likelihood detection; random processes; signal sampling; CUSUM test; GLR; MLE; abrupt changes detection; fixed sample size; maximum likelihood detector; mean; multivariate Gaussian random signal; optimum sequential tests; quadratic tests; sequential detector; Change detection algorithms; Delay effects; Detection algorithms; Ellipsoids; Gaussian distribution; Probability; Sequential analysis; Signal to noise ratio; Testing; Vectors;
Journal_Title :
Signal Processing, IEEE Transactions on
