On the Use of Higher-Order Statistical Tests in the Analysis of Time Series Associated with Space Data. Utilisation de Tests Basés sur des Statistiques D’Ordre Supérieur dans L’Analyse de Séries Temporelles Mesurées dans L’Espace

On the Use of Higher-Order Statistical Tests in the Analysis of Time Series Associated with Space Data.

Utilisation de Tests Basés sur des Statistiques D’Ordre Supérieur dans L’Analyse de Séries Temporelles Mesurées dans L’Espace

A. Masson B.B. Shishkov  F. Lefeuvre 

Laboratoire de Physique et Chimie de l’Environnement, Centre National de la Recherche Scientifique, 45071 Orléans, Cedex 2, France

Page: 
59-78
|
Received: 
5 March 1999
| |
Accepted: 
N/A
| | Citation

OPEN ACCESS

Abstract: 

Tests of hypotheses based on Higher Order Statistics (HOS) are reviewed in the particular context of the identification of nonlinear processes in space plasma. The time series under study are associated with the measurements of electric or/and magnetic field components, or/and counting rates of particles. The basic principles of HOS techniques are reviewed. A general and unified procedure is suggested in order to construct statistical tests: (1) for detecting a non-gaussian or transient signal in a gaussian (or non-gaussian) noise, (2) testing a stochastic time series for non-gaussianity (including non-linearity), (3) studying non-linear wave interactions by using the kth-order coherency function. Asymptotic theory of estimates of the kthorder spectra is implemented in a digital signal processing framework. The effectiveness of the signal detection algorithms is demonstrated through computer simulations. Examples of application on the analysis of satellite data are given.

Résumé

Des tests d’hypothèses basés sur des statistiques d’ordre supérieur sont revus dans le contexte particulier de l’identification de processus non-linéaires dans les plasmas spatiaux. Les séries temporelles étudiées sont associées à la mesure de composantes du champ électrique et/ou magnétique d’ondes ou de turbulences, et/ou de données particules. Les principes de base des statistiques d’ordre supérieur sont brièvement rappelés. Une procédure générale et unifiée est suggérée afin de construire des tests statistiques permettant : (1) de détecter des signaux non-gaussiens ou transitoires au sein d’un bruit gaussien (ou non-gaussien), (2) de tester si une série temporelle est associée ou non à un processus stochastique issu d’un processus non-linéaire, (3) d’étudier des interactions non-linéaires à plusieurs ondes par l’utilisation de la fonction de cohérence d’ordre k. La théorie asymptotique des estimés des spectres d’ordre k est mise en œuvre dans le cas discret. L’efficacité des algorithmes de détection est démontrée par le biais de simulations numériques. Des exemples d’applications à des données satellites sont présentés.

Keywords: 

Higher-order statistics; non-linear phenomena; transient phenomena; space plasmas.

Mots clés

Statistiques d’ordre supérieur ; phénomènes non linéaires ; phénomènes transitoires ; plasmas spatiaux.

1. Introduction
2. General Background
3. Statistical Tests Based on HOS: Detection Scheme and HOS Estimators
4. Detection of Nongaussian Signals by Third-Order Spectral Analysis
5. Testing for Linearity of Stationary Time Series
6. Detecting a Transient Signal by Bispectral Analysis
7. Applications of the HOS Technique to Nonlinear Wave-Wave Interactions
8. Applications to Satellite Data
9. Conclusion
Notations and abbreviations
Appendix
  References

[1] T. W. Anderson, An Introduction to Multivariate Statistical Analysis New York: Wiley 1958. 

[2] H. Akaike, “Note on higher-order spectra”, Ann. Inst. Statist. Math., vol. 18, pp. 123-126,1966. 

[3] P. O. Ambard, J. L. Lacoume, and J. M. Brossier, “Transient detection, higher-order time-frequency distributions and the entropy”, in Proc. IEEE Work on Higher-Order Spectral Analysis (South Lake Tahoe, CA, June 7-9, 1993), pp. 265-269. 

[4] A. Blanc-Lapierre and R. Fortet, Théorie des Fonctions Aléatoires. Paris: Masson, 1953. 

[5] D. R. Brillinger, “An introduction to polyspectra”, Ann. Math. Statist., vol. 36, pp. 1351-1374, 1965. 

[6] D. R. Brillinger and M. Rosenblatt, “Asymptotic theory of estimates of kth order spectra”, in Spectral Analysis of Time Series, pp. 135-188, B. Harris, Ed. New York: Wiley, 1967. 

[7] D. R. Brillinger, Time Series, Data Analysis and Theory, expanded. New York: Holt, Rine-hart and Winston, 1981. 

[8] V. Chandran and S. Elgar, “A general procedure for the derivation of  principal domains of higher-order spectra”, IEEE Trans. Signal Process., vol. 42, n° 1, pp. 229-233, Jan. 1994. 

[9] R. F. Dwyer, “Detection of non-Gaussian signals by frequency domain kurtosis estimation”, in Proc. 1983 IEEE Int. Conf. on Acoustics, Speech, and Signal Processing, pp. 607-610, 1983. 

[10] T. Dudok de Wit and V. V. Krasnosel’skikh, “Wavelet bicoherence analysis of strong plasma turbulence at the Earth’s quasi-parallel bowshock”, Phys. Plasma, 2, 262-273, 1995. 

[11] S. Elgar, “Relationships involving third moments and bispectra of a harmonic process”, IEEE , vol. ASSP-35, pp. 1725-1726, 1987. 

[12] M. M. Gabr and T. Subba Rao, “On the estimation of bispectral density function in the case of randomly missing observations”, IEEE Trans. Signal Processing, vol. 42, n° 1, pp. 211-216, Jan. 1994. 

[13] L. M. Garth and H. Vincent Poor, “Detection of non-Gaussian signals: a paradigm for modern statistical signal processing”, Proc. of the IEEE, vol. 82, n° 7, pp. 1060-1095, July 1994.

[14] G. B. Gianakis and M. K. Tsatsanis, “Signal detection and classification using matched filtering and higher-order statistics”, IEEE Trans. Acoust. Speech, and Signal Processing, vol. 38, n° 7, pp. 1284-1296, July 1990. 

[15] G. B. Gianakis and M. K. Tsatsanis, “A unifying maximum-likelihood view of cumulant and polyspectral measures for non-gaussian signal classification and estimation”, “Signal detection and classification using matched filtering and higher-order statistics”, IEEE Trans. Inform. Theory, vol. 38, n° 7, pp. 1284-1296, 1990. 

[16] M. J. Hinich and C. S. Clay, “The application of the discrete Fourier transform in the estimation of power spectra, coherence, and bispectra of geophysical data”, Rev. Geophys., vol. 6, pp. 347-363, 1968. 

[17] M. J. Hinich, “Testing for Gaussianity and linearity of a stationary time series”, J. Time Series Anal., vol. 3, pp. 169-176, 1982. 

[18] M. J. Hinich and G. R. Wilson, “Detection of non-Gaussian signals in nonGaussian noise using the bispectrum”, IEEE Trans. Acoust., Speech, and Signal Process., vol. 38, n° 7, pp. 1126-1131, July 1990. 

[19] M. J. Hinich, “Detecting a transient signal by bispectral analysis”, IEEE Trans. Acoust, Speech, and Signal Process., vol. 38, n° 7, pp. 1277-1283, July 1990. 

[20] P. J. Huber, B. Kleiner, T. Gasser, and G. Dumermuth, “Statistical methods for investigating phase relations in stationary stochastic processes”, IEEE Trans. Audio Electro-acoustics, AV-19, pp. 78-86, 1971. 

[21] M. G. Kendall and A. Stuart, The Advanced Theory of Statistics, vol. 1, 3rd ed. New York: Hafner, 1958. 

[22] D. Kletter and H. Messer, “Suboptimal detection of non-Gaussian signals by third-order spectral analysis”, IEEE, Trans. Acoust., Speech, and Signal Process., vol. 38, n° 6, pp. 901-909, June 1990. 

[23] Y. C. Kim and E. J. Powers, “Digital bispectral analysis of self-excited fluctuation spectra”, Phys. Fluids, 21(8), pp. 1452-1453, August, 1978. 

[24] Y. C. Kim and  E. J. Powers, “Digital bispectral analysis  and  its applications to nonlinear wave interactions”, IEEE Trans. Plasma Science, vol. PS7, n° 2, pp. 120-131, June 1979. 

[25] K. I. Kim and E. J. Powers, “A digital method of modeling quadratically nonlinear systems with a general random input”, IEEE Trans. Acoust., Speech, and Signal Process., vol. 36, n° 11, pp. 1758-1769, November 1988. 

[26] V. Kravtchenko-Berejnoi, F. Lefeuvre, V. Krasnossel’skikh, D. Lagoutte, “On the use of tricoherent analysis to detect non-linear wave-wave interactions”, Signal Processing, 42, pp. 291-309, 1995. 

[27] J.-L. Lacoume, P.-O. Amblard, and P. Comon, Statistiques d’ordre supérieur pour le traitement du signal. Paris: Masson, 1997. 

[28] D. Lagoutte, F. Lefeuvre, and J. Hanasz, “Application of bicoherence analysis in the study of wave interactions in space plasma”, J. Geophys. Res., 94, 435, 1989. 

[29] D. Lagoutte and P. Latremoliere, SWAN, Software for Waveform Analysis, Version 2.0, LPCE-CNRS, May 1997. 

[30] K. S. Lii and M. Rosenblatt, “Deconvolution and estimation of transfer function phase and coefficients for non-Gaussian linear processes”, Ann. Statist., vol. 10, n° 4, pp. 1195-1208, 1982. 

[31] G. Mann, H. Lühr, W. Baumjohann, “Statistical analysis of short largeamplitude magnetic field structures in the vicinity of the quasi-parallel bow shock”, J. Geophys. Res., Vol. 99, n° A7, pp. 13,315, 1994. 

[32] A. Masson, F. Lefeuvre, Z. Y. Zhao, D. Lagoutte, J. L. Rauch, “Observation of nonlinear interactions in large scale density enhancements of the highlatitude ionosphere”, J. Geophys. Res. Vol. 104, N° A10, pp. 22,499, 1999. 

[33] J. M. Mendel, “Tutorial on higher-order statistics (spectra) in signal processing  and system theory: theoretical results and some applications”, Proc. of the IEEE, vol. 79, n° 3, pp. 278-305, March 1991.

[34] C. L. Nikias and M. S. Raghuveer, “Bispectrum estimation: a digital signal processing framework”, Proc. of the IEEE, vol. 75, n° 7, pp. 869-891, July 1987. 

[35] C. L. Nikias and A. P. Petropulu, Higher-Order Spectra Analysis: A Nonlinear Signal Processing Framework. Englewood Cliffs, NJ: PrenticeHall, 1993. 

[36] L. A. Pflug, G. E. Ioup, J. W. Ioup, R. L. Field, “Properties of higher-order correlations and spectra for bandlimited deterministic transients”, J. Acoust. Soc. Am. 91(2), pp. 975-988, February 1992.

[37] L. A. Pflug, G. E. Ioup, J. W. Ioup, K. H. Barues, R. L. Field, G.H. Rayborn, “Detection of oscillatory and impulsive transients using higher-order correlations and spectra”, J. Acoust. Soc. Am. 91(5), pp. 2763-2776, May 1992. 

[38] L. A. Pflug, G. E. Ioup, J. W. Ioup, R. L. Field, “Prediction of Signal-tonoise ratio gain for passive higher-order correlation detection of energy transients”, J. Acoust. Soc. Am. 98(1), pp. 249-260, July 1995. 

[39] O. Randriamboarison, V. Kravtchenko-Berejnoi, D. Lagoutte, F. Lefeuvre, “Analysis of models of nonlinear interactions by higher-order statistics”, in Proc. Spatio-Temporal Analysis for Resolving Plasma Turbulence, pp. 285288, Aussois, France, 1993. 

[40] M. Rosenblatt and J. W. Van Ness, “Estimation of the bispectrum”, Ann. Math. Statist., 36, pp. 1120-1136, 1965. 

[41] M. Rosenblatt, Stationary Sequences and Random Fields, Boston: Birkhauser, 1985. 

[42] G. G. Roussas, Contiguity of Probability Measures. Cambridge University Press, 1972. 

[43] S. J. Schwartz, D. Burgess, W. P. Wilkinson, R. L. Kessel, M. Dunlop, and H. Lühr, “Observations of Short Large-Amplitude Magnetic Structures at a Quasi-Parallel Shock”, J. Geophys. Res., Vol. 97, n° A4, pp. 4209, 1992. 

[44] B. B. Shishkov, “Locally asymptotic normality for the parametric models of time series in the spectral domain”, Compt. Rend. Acad. Bulg. Sc., vol. 43, pp. 17-20, 1990. 

[45] B. B. Shishkov, “Testing hypotheses on distribution’s parameters of time series and applications to signal detection”. Radio Eng. Electron. Phys., vol. 39, pp. 411-423, 1993. 

[46] B. B. Shishkov, H. Tsuji, J. Xin, Y. Hase, “Spatial and temporal processing of cyclostationary signals in array antennas based on linear prediction model”, Submission to Telecommunications Advancement Organization (TAO) of Japan, December, 1997. 

[47] A. N. Shiryaev, “Some problems in the spectral theory of higher-order moments”, Theor. Prob. Appl., vol. 5, pp. 265-284, 1960. 

[48] A. D. Spaulding, “Locally optimum and suboptimum detector performance in a non-Gaussian interference environment”, IEEE Trans. Commun., vol. COM-33, n° 6, pp. 509-517, 1985. 

[49] T. Subba Rao and M. M. Gabr, “A test for linearity of stationary time series”, J. Time Series Anal., vol. 1, n° 2, pp. 145-158, 1980. 

[50] T. Subba Rao and M. M. Gabr, “An introduction to bispectral analysis and bilinear time series model”, in Lecture Notes in Statistics, D. Brillinger et al., Eds. New York: Springer-Verlag, 1984.

[51] C. W. Therrien, Discrete Random Signals and Statistical Signals Processing. Prentice Hall, Englewood Cliffs, NJ, 1992. 

[52] V. Y. Trakhtengerts and M. Hayakawa, “A wave interaction in whistler frequency range in space plasma”, J. Geophys. Res., 98, 19,205, 1993. 

[53] J. W. Van Ness, “Asymptotic normality of bispectral estimates”, Ann. Math. Statist., vol. 37, pp. 1257-1275, 1966. 

[54] Workshop on Turbulence in Space Plasmas. Non-Linear Processes and Data Filtering Techniques, Garchy (France), 5-7 May 1998.