Least Squares and Maximum Likelihood Estimation of Mixed Spectra

Brynolfsson, JohanLund University,Lunds universitet,Statistical Signal Processing Group,Forskargrupper vid Lunds universitet,Matematisk statistik,Matematikcentrum,Institutioner vid LTH,Lunds Tekniska Högskola,Lund University Research Groups,Mathematical Statistics,Centre for Mathematical Sciences,Departments at LTH,Faculty of Engineering, LTH (författare)

In this paper, we propose a novel 1-D spectralestimator for signals with mixed spectra. The proposed methodis partly based on the recently introduced smooth spectralestimator LIMES, in which the smoothness is accounted for byassuming linearity within predefined segments of the spectrum.The proposed method utilizes this formulation but also allowssegments to change size to better estimate the spectrum, therebyallowing for the estimation of spectra that are neither completelysmooth or sparse in frequency, but rather contains a mixtureof such components. Using simulated data, we illustrate theperformance of the proposed estimator, comparing to other recentspectral estimation techniques.

Swärd, JohanLund University,Lunds universitet,Statistical Signal Processing Group,Forskargrupper vid Lunds universitet,Matematisk statistik,Matematikcentrum,Institutioner vid LTH,Lunds Tekniska Högskola,Lund University Research Groups,Mathematical Statistics,Centre for Mathematical Sciences,Departments at LTH,Faculty of Engineering, LTH(Swepub:lu)mats-jsd (författare)
Jakobsson, AndreasLund University,Lunds universitet,Biomedical Modelling and Computation,Forskargrupper vid Lunds universitet,Statistical Signal Processing Group,Matematisk statistik,Matematikcentrum,Institutioner vid LTH,Lunds Tekniska Högskola,Lund University Research Groups,Mathematical Statistics,Centre for Mathematical Sciences,Departments at LTH,Faculty of Engineering, LTH(Swepub:lu)mats-aaj (författare)
Sandsten, MariaLund University,Lunds universitet,Statistical Signal Processing Group,Forskargrupper vid Lunds universitet,Matematisk statistik,Matematikcentrum,Institutioner vid LTH,Lunds Tekniska Högskola,Lund University Research Groups,Mathematical Statistics,Centre for Mathematical Sciences,Departments at LTH,Faculty of Engineering, LTH(Swepub:lu)tde-mha (författare)
Statistical Signal Processing GroupForskargrupper vid Lunds universitet (creator_code:org_t)

Ingår i:26th European Signal Processing Conference, EUSIPCO 2018., s. 2345-23499789082797015

Av författaren/redakt...: Brynolfsson, Joh ...; Swärd, Johan; Jakobsson, Andre ...; Sandsten, Maria

Om ämnet

Kungliga biblioteket hanterar dina personuppgifter i enlighet med EU:s dataskyddsförordning (2018), GDPR. Läs mer om hur det funkar här.
Så här hanterar KB dina uppgifter vid användning av denna tjänst.