Computer Science – Information Theory
Scientific paper
2011-06-06
Computer Science
Information Theory
8 pages, 2 figures, submitted to the IEEE Transactions on Information Theory
Scientific paper
We examine codes, over the additive Gaussian noise channel, attaining a minimum required rate at some specific signel-to-noise ratio (snr). For such codes we provide a tight lower bound on the minimum mean square error (MMSE), valid for any snr. This minimum required rate is below capacity, and hence these are non-optimal codes (alternatively referred to as "bad" codes). We show that the lower bound is attainable by superposition codebooks that are optimal for a specific degraded Gaussian broadcast channel (BC). It is also proved that the MMSE function of codes, attaining a minimum required rate at some snr, and the lower bound on the MMSE at some other snr, is completely defined for all snr, and is the one obtained by the corresponding superposition codebooks.
Bustin Ronit
Shamai Shlomo
No associations
LandOfFree
MMSE of "Bad" Codes does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with MMSE of "Bad" Codes, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and MMSE of "Bad" Codes will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-389524