On ordered normally distributed vector parameter estimates

The ordered values of a sample of observations are called the order statistics of the sample and are among the most important functions of a set of random variables in probability and statistics. However the study of ordered estimates seems to have been overlooked in maximum-likelihood estimation. Therefore it is the aim of this communication to give an insight into the relevance of order statistics in maximum-likelihood estimation by providing a second-order statistical prediction of ordered normally distributed estimates. Indeed, this second-order statistical prediction allows to refine the asymptotic performance analysis of the mean square error (MSE) of maximum likelihood estimators (MLEs) of a subset of the parameters. A closer look to the bivariate case highlights the possible impact of estimates ordering on MSE, impact which is not negligible in (very) high resolution scenarios. HighlightsWe provide a second-order statistical prediction of ordered normal estimates.It allows to refine the asymptotic performance analysis of the MSE of MLEs.The impact of estimates ordering is highlighted for two closely spaced sources.

[1]  Ingram Olkin,et al.  Correlation Analysis of Extreme Observations from a Multivariate Normal Distribution , 1995 .

[2]  H. M. Barakat.  Multivariate order statistics based on dependent and nonidentically distributed random variables , 2009, J. Multivar. Anal..

[3]  A. Genz.  Numerical Computation of Multivariate Normal Probabilities , 1992 .

[4]  P. Larzabal,et al.  On the High-SNR Conditional Maximum-Likelihood Estimator Full Statistical Characterization , 2006, IEEE Transactions on Signal Processing.

[5]  Jian Li,et al.  Maximum likelihood angle estimation for signals with known waveforms , 1993, IEEE Trans. Signal Process..

[6]  Philippe Loubaton,et al.  On the resolution probability of conditional and unconditional maximum likelihood DOA estimation , 2013, 21st European Signal Processing Conference (EUSIPCO 2013).

[7]  Björn E. Ottersten,et al.  Sensor array processing based on subspace fitting , 1991, IEEE Trans. Signal Process..

[8]  M. Genton,et al.  On the exact distribution of linear combinations of order statistics from dependent random variables , 2007 .

[9]  Petre Stoica,et al.  Performance study of conditional and unconditional direction-of-arrival estimation , 1990, IEEE Trans. Acoust. Speech Signal Process..

[10]  M. P. Clark.  On the resolvability of normally distributed vector parameter estimates , 1995, IEEE Trans. Signal Process..

[11]  J. Hosking.  L‐Moments: Analysis and Estimation of Distributions Using Linear Combinations of Order Statistics , 1990 .

[12]  T. Sargent,et al.  The multivariate normal distribution , 1989 .

[13]  Nicola Loperfido,et al.  A note on skew-elliptical distributions and linear functions of order statistics , 2008 .

[14]  Bjorn Ottersten,et al.  Exact and Large Sample ML Techniques for Parameter Estimation and Detection in Array Processing , 1993 .

[15]  Eric Chaumette,et al.  On the Accuracy and Resolvability of Vector Parameter Estimates , 2014, IEEE Transactions on Signal Processing.