We consider the transmission of two correlated Gaussian sources over a Gaussian weak interference channel. Each transmitter has access to one component of a bivariate Gaussian source, and each of these components need to be reconstructed at the corresponding receiver, under the squared-error distortion measure. We are interested in characterizing the region of average distortion pairs achievable simultaneously at the two receivers. We derive an outer bound on the achievable region, and show that, under certain conditions, achievable distortion pair with uncoded transmission lies on this outer bound; therefore, it is optimal. In particular, optimality of uncoded transmission is shown to hold for all signal-to-noise ratio values below a certain threshold which depends on the correlation coefficient between the sources.
Correlated Gaussian sources over Gaussian weak interference channels / Aguerri, I. E.; Gunduz, D.. - (2015), pp. 84-88. (Intervento presentato al convegno IEEE Information Theory Workshop, ITW 2015 tenutosi a Lotte City Hotel Jeju, kor nel 2015) [10.1109/ITWF.2015.7360739].
Correlated Gaussian sources over Gaussian weak interference channels
D. Gunduz
2015
Abstract
We consider the transmission of two correlated Gaussian sources over a Gaussian weak interference channel. Each transmitter has access to one component of a bivariate Gaussian source, and each of these components need to be reconstructed at the corresponding receiver, under the squared-error distortion measure. We are interested in characterizing the region of average distortion pairs achievable simultaneously at the two receivers. We derive an outer bound on the achievable region, and show that, under certain conditions, achievable distortion pair with uncoded transmission lies on this outer bound; therefore, it is optimal. In particular, optimality of uncoded transmission is shown to hold for all signal-to-noise ratio values below a certain threshold which depends on the correlation coefficient between the sources.
Pubblicazioni consigliate
I metadati presenti in IRIS UNIMORE sono rilasciati con licenza Creative Commons CC0 1.0 Universal, mentre i file delle pubblicazioni sono rilasciati con licenza Attribuzione 4.0 Internazionale (CC BY 4.0), salvo diversa indicazione.
In caso di violazione di copyright, contattare Supporto Iris
