The Achievable Distortion Region of Sending a Bivariate Gaussian Source on the Gaussian Broadcast Channel

We provide a complete characterization of the achievable distortion region for the problem of sending a bivariate Gaussian source over bandwidth-matched Gaussian broadcast channels, where each receiver is interested in only one component of the source. This setting naturally generalizes the simple single Gaussian source bandwidth-matched broadcast problem for which the uncoded scheme is known to be optimal. We show that a hybrid scheme can achieve the optimum for the bivariate case, but neither an uncoded scheme alone nor a separation-based scheme alone is sufficient. We further show that in this joint source channel coding setting, the Gaussian scenario is the worst scenario among the sources and channel noises with the same covariances.

[1]  Suhas N. Diggavi,et al.  Approximate characterizations for the Gaussian broadcasting distortion region , 2009, 2009 IEEE International Symposium on Information Theory.

[2]  Thomas J. Goblick,et al.  Theoretical limitations on the transmission of data from analog sources , 1965, IEEE Trans. Inf. Theory.

[3]  Sang Joon Kim,et al.  A Mathematical Theory of Communication , 2006 .

[4]  Suhas N. Diggavi,et al.  Approximate Characterizations for the Gaussian Source Broadcast Distortion Region , 2011, IEEE Transactions on Information Theory.

[5]  Amos Lapidoth,et al.  Broadcasting Correlated Gaussians , 2007, IEEE Transactions on Information Theory.

[6]  Thomas M. Cover,et al.  Elements of Information Theory , 2005 .

[7]  Meir Feder,et al.  Distortion Bounds for Broadcasting With Bandwidth Expansion , 2006, IEEE Transactions on Information Theory.

[8]  Ertem Tuncel,et al.  Separate source-channel coding for broadcasting correlated Gaussians , 2011, 2011 IEEE International Symposium on Information Theory Proceedings.

[9]  Sriram Vishwanath,et al.  Hybrid coding for Gaussian broadcast channels with Gaussian sources , 2009, 2009 IEEE International Symposium on Information Theory.

[10]  Amos Lapidoth,et al.  Sending a Bivariate Gaussian Over a Gaussian MAC , 2010, IEEE Transactions on Information Theory.