M. Basseville and A. Benveniste, editors. Detection of Abrupt Changes in Signals and Dynamical Systems. Springer-Verlag, 1986.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....with normal behavior of the system. The passive approach is used to continuously monitor the system in particular when the detector has no way of acting upon the system, for material or security reasons. Most of the work in the area of failure detection is geared towards this type of approach [1], 12] 15] The active approach to failure detection consists in acting upon the system on a periodic basis or at critical times using a test signal, which we call an auxiliary signal, in order to exhibit abnormal behaviors which would otherwise remain undetected during normal operation. The ....

Basseville, M. and Benveniste, A., (eds), Detection of abrupt changes in signals and dynamical systems, Lecture notes in Control and Information Science, vol. 77, Springer-Verlag, 1985.

....model for the measurements and three different change detection algorithms. The change detection algorithms are similar to the one discussed here. Both our detector and the work of Andre Obrecht is based on a number of papers by Basseville and Benveniste [Basseville and Benveniste, 1983, Basseville, 1986, Basseville et al. 1986, Basseville et al. 1987, Basseville, 1988, Benveniste et al. 1987] These works have been applied to segmentation of EEG, ECG, speech, and geophysical signals. The best reference is the collection of papers in [Basseville and Benveniste, 1986] Their works are all ....

....the measurements and three different change detection algorithms. The change detection algorithms are similar to the one discussed here. Both our detector and the work of Andre Obrecht is based on a number of papers by Basseville and Benveniste [Basseville and Benveniste, 1983, Basseville, 1986, Basseville et al. 1986, Basseville et al. 1987, Basseville, 1988, Benveniste et al. 1987] These works have been applied to segmentation of EEG, ECG, speech, and geophysical signals. The best reference is the collection of papers in [Basseville and Benveniste, 1986] Their works are all related to the sequential ....

[Article contains additional citation context not shown here]

M. Basseville and A. Benveniste, editors. Detection of Abrupt Changes in Signals and Dynamical Systems. Springer-Verlag, 1986.

....Tests, Change Point, Periodogram, Tapered Data, Stationary Process. Mathematics subject classi cations (1991) 62G10, 62G20, 62M10, 62M15. 1 Introduction The problem of detecting a change point in the properties of a process has been extensively studied, see for a general survey the books [1], 2] and [7] or more recently the works of [21] or [28] The change point problem can be formulated in two di erent ways: a sequential problem called online and another one a posteriory called o line . We consider here this last case where one has to decide between homogeneity and ....

M. Basseville and A. Benveniste, editors. Detection of abrupt changes in signals and dynamical systems, volume 77. Springer, 1986.

.... based on a likelihood ratio test such as described by Van Trees [3, chap.2] As indicated above, we are interested in determining the time at which the process z t in (3) changes abruptly from zero to one; the detection of such abrupt changes is discussed in the texts by Basseville and Benveniste [4] and Basseville and Nikiforov [5] We therefore define the likelihood ratio L as L = P r[T tjobservations of y from 0 to t] P r[T tjobservations of y from 0 to t] and use (6) with z replaced by z to obtain L = P r[T tjy s ; 0 s t] P r[T tjy s ; 0 s t] z 1 Gamma z : 10) ....

M. Basseville and A. Benveniste, Detection of Abrupt Changes in Signals and Dynamical Systems, SpringerVerlag, 1980.

....examples to illustrate the superiority of the new algorithm over more conventional techniques. 1 Introduction The detection of changes in the dynamics of systems from noisy measurements has been well studied and is comprehensively documented in the survey chapters and books [19] 10] 1] and [2]. Two main approaches that have sprung from this work are Kalman Filter based approaches introduced by Mehra and Peschon [15] and likelihood ratio test approaches. In the spirit of these model based solutions, we propose in this chapter a likelihood ratio test based on the best linear unbiased ....

....generated the data fY 0 ; Gamma 0 g. That is, we wish to decide between the hypotheses H 0 and H 1 : H 0 : fi 0 = 0 H 1 : fi 0 6= 0 where fi 0 , fi 0 Gamma fi F . For this decision problem, there are a number of possible decision strategies that could be applied. Many of these [9] [2] depend upon testing the logarithm of the likelihood ratio: fi 0 ) ln ae P(Y jH 1 ) P(Y jH 0 ) oe (9) against a threshold . In order to calculate the test statistic (9) we use the fact that with the Gaussian assumptions the difference in the parameter estimates fi , fi 0 ....

M. Basseville and A. Benveniste. Detection of Abrupt Changes in Signals and Dynamical Systems. Springer-Verlag, 1986.

....time. Two types of problems can be considered. One is the detection of disorder (hypotheses testing) or quickest detection of disorder (sequential hypotheses testing) The second is the estimation of the change time (parameter estimation) Applications can be found in Basseville and Benveniste [1]. The quickest detection problem is considered in Shiryayev [6] for observations of the form dX t = a 1 (t ) dt oe dW t ; X 0 = 0 ; t 0 ; where is an exponential random variable on [0; 1) and the parameters a and oe are known. In this Bayesian framework, an optimal stopping time is ....

M. BASSEVILLE and A. BENVENISTE, editors. Detection of Abrupt Changes in Signals and Dynamical Systems, volume 77 of Lecture Notes in Control and Information Sciences. Springer Verlag, 1986.

....be compared to the quadratic in time complexity of the exact GLR. I. Introduction The problem of detecting abrupt changes in linear systems and signals occurs in many applications. The practical and theoretical interest in this field are reflected in a large number of surveys, for instance [2] [3], 4] 6] 10] 11] 14] One of the most powerful methods in change detection is the generalized likelihood ratio (GLR) test proposed in [15] It applies to cases of abrupt changes in the state of an arbitrary linear system with known dynamics. Its general applicability has contributed to ....

M. Basseville and A. Benveniste. Detection of Abrupt Changes in Signals and Dynamical Systems. Lecture Notes in Control and Information Sciences. Springer-Verlag, 1986.

....serait de remplacer (dans le cas du signal HF) l estimation d un mod ele gaussien par un mod ele de type 2 . Il reste aussi a evaluer, concurremment au mod ele des chaines de Markov cach ees qui a et e etudi e ici, des m ethodes autor egressives classiques de d etection de rupture (cf. BAB 86] Enfin une validation de la m ethode par des exp erimentations pharmacologiques animales (il existe des m edicaments qui bloquent sp ecifiquement une branche ou l autre du syst eme nerveux autonome) semble n ecessaire et reste a faire. ....

BASSEVILLE, M. et BENVENISTE, A. (Eds.) : Detection of abrupt changes in signals and dynamical systems. Lecture Notes in Control and Information Sciences, 77. Springer Verlag, New-York, 1986.

....or innovations in a Kalman filter, that is, the difference between the observations and predicted observations based on model and data, represent a statistically whitened version of the observations resulting from what is in essence a high pass filter. As discussed in many papers and books ([4, 51], for example) discontinuities in the data being processed then lead to distinctive signatures which can be looked for using optimal detection methods. In a similar fashion we can compute the residuals of the MR estimates: s) y(s) Gamma C(s)b x s (s) 63) for the chopper sequence, an image ....

....signature, as the helicopter rotors, nearly imperceptible in Figure 26 are clearly in evidence in Figure 31 because of the motion discontinuity. As we have indicated, statistically optimal methods for using residuals analogous to these have been developed for time series, and, as discussed in [4, 51], such methods require error covariance information from the estimator in order to specify the optimal detection procedure. Since the MR algorithm also produces such error covariance information it is possible to develop optimal detection methods in this imaging context as well. Such a method is ....

M. Basseville and A. Benveniste. Detection of Abrupt Changes in Signals and Dynamical Systems, volume 77 of Lecture Notes in Control and Information Sciences. Springer-Verlag, 1986.

....reasons. First, Gauss Markov processes (and their non Gaussian generalizations) are excellent models for a wide class of interesting problems and phenomena, including many arising in the design and control of dynamic systems [3, 16, 69] biomedical, seismic and geophysical signal processing [6, 57], and speech and image processing [108, 111, 88] Second, their simple structure leads naturally t t 1 Figure 1 5: A first order tree, corresponding to the set of integers, is shown. Time recursive models defined on the first order tree, such as (1.19) provide a rich modeling framework and lead ....

....or innovations in a Kalman filter, that is, the difference between the observations and predicted observations based on model and data, represent a statistically whitened version of the observations resulting from what is in essence a high pass filter. As discussed in many papers and books ([6, 145], for example) discontinuities in the data being processed then lead to distinctive signatures which can be looked for using Figure 2 29: The smoothing filter residuals shown above can be used to develop adaptive algorithms for the motion based object detection. optimal detection methods. In a ....

[Article contains additional citation context not shown here]

M. Basseville and A. Benveniste. Detection of Abrupt Changes in Signals and Dynamical Systems, volume 77 of Lecture Notes in Control and Information Sciences. Springer-Verlag, 1986.

....and NSF grant MIP 9015281. called knots. Spline approximation is both a nice theoretical problem and has many applications. Some examples are initial value problems [3] nonlinear boundary value problems [3] eigenvalue problems [3] and detection and estimation of abrupt changes in signals [4] [5]. The latter problem is the one in which we are interested. Namely, we would like to approximate a given function by a spline, placing the knots at the points of abrupt changes of the function and approximating the pieces between these points by polynomials. As the name detection of abrupt ....

Michele Basseville and Albert Benveniste, editors. Detection of Abrupt Changes in Signals and Dynamical Systems. Springer-Verlag, 1986.

No context found.

M. Basseville and A. Benveniste, editors. Detection of Abrupt Changes in Signals and Dynamical Systems. Springer-Verlag, 1986.

No context found.

Basseville M. and A. Benveniste. Detection of abrupt changes in signals and dynamical systems. Springer, Berlin, 1986.

No context found.

Basseville M. and Benveniste A. (eds.) (1986), Detection of Abrupt Changes in Signals and Dynamical Systems, Springer-Verlag, Berlin.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC