A weighted mixed statistics algorithm for blind source separation

Klajman,; Constantinides,

doi:10.1109/icassp.2002.1004861

Cited by 2 publications

(4 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Let x z be the unique positive zero argument of F (1) . Since F (1) is negative in a neighborhood of zero, and F (1) has no zeros for x ≤ x z , F (1) is negative in the interval [0,…”

Section: Application Of Hurwitz Theoremmentioning

confidence: 99%

“…Since F (1) is positive for large value of x, F (1) must change its sign at x z , and hence it is positive in the interval…”

Section: Application Of Hurwitz Theoremmentioning

confidence: 99%

“…For all integers m, we note that f and g m have the same number of zeros. Using Rolles theorem, it can be proved that the derivative of g m which we denote g (1) m and which is given by:…”

Section: Appendix C Proof Of Theoremmentioning

confidence: 99%

“…Many parameter estimation techniques use combined criteria to exploit different features or properties of the considered signals and hence improve the estimation performance. Examples of such combined techniques include the blind source separation (BSS) method in [1], and the blind equalization method in [2] where second and higher order statistics based criteria are combined to restore the source signals, and the channel equalization and offset estimation technique in [3], where again two criteria based on two different features of the transmit signals are jointly used to improve the receiver performance.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Quasi-Convexity of the Asymptotic Channel MSE in Regularized Semi Blind Estimation

Kammoun

Abed-Meraim

Affes

2011

IEEE Trans. Inform. Theory

View full text Add to dashboard Cite

In this paper, the quasi-convexity of a sum of quadratic fractions in the form n i=11+c i x 2 (1+d i x) 2 is demonstrated where ci and di are strictly positive scalars, when defined on the positive real axis R + . It will be shown that this quasi-convexity guarantees it has a unique local (and hence global) minimum.Indeed, this problem arises when considering the optimization of the weighting coefficient in regularized semi-blind channel identification problem, and more generally, is of interest in other contexts where we combine two different estimation criteria.Note that V. Buchoux et.al have noticed by simulations that the considered function has no local minima except its unique global minimum but this is the first time this result, as well as the quasi-convexity of the function is proved theoretically.

show abstract

“…Let x z be the unique positive zero argument of F (1) . Since F (1) is negative in a neighborhood of zero, and F (1) has no zeros for x ≤ x z , F (1) is negative in the interval [0,…”

Section: Application Of Hurwitz Theoremmentioning

confidence: 99%

“…Since F (1) is positive for large value of x, F (1) must change its sign at x z , and hence it is positive in the interval…”

Section: Application Of Hurwitz Theoremmentioning

confidence: 99%

“…For all integers m, we note that f and g m have the same number of zeros. Using Rolles theorem, it can be proved that the derivative of g m which we denote g (1) m and which is given by:…”

Section: Appendix C Proof Of Theoremmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations