Equivocation limits in Gaussian additive channels
Abstract
Equivocation in a Gaussian additive channel is defined as H(X/X + Z(subt)) or as the conditional entropy of X entry and X + Z (subt) exit of the channel. Equivocation is obviously a function of t power of additive Gaussian interference Z(subt). The upper and lower limits of equivocation as the function of interference power t are investigated. The lower limit passes the concavity of entropy power in the function of additional interference power. This limit has numerous uses in information theory: it permits establishment of the optimality of two extreme points of the general Gaussian interference channel. The upper limit is a consequence of the inequality of Shannon entropy power. This limit establishes the fact that the very rapid increase in equivocation is logarithmic.
 Publication:

NASA STI/Recon Technical Report N
 Pub Date:
 July 1983
 Bibcode:
 1983STIN...8513137C
 Keywords:

 Communication;
 Concavity;
 Gauss Equation;
 Information Theory;
 Convolution Integrals;
 Entropy (Statistics);
 Vectors (Mathematics);
 Communications and Radar