In this paper we consider point-to-point multiantenna channels with certainblock distributional symmetries which do not require the entries of the channelmatrix to be either Gaussian, or independent, or identically distributed. Amain contribution is a capacity theorem for these channels, which might beregarded as a generalization of Telatar's theorem (1999), which reduces thenumerical optimization domain in the capacity computation. With thisinformation theoretic result and some free probability arguments, we prove anasymptotic capacity theorem that, in addition to reducing the optimizationdomain, does not depend on the dimension of the channel matrix. This theoremallows us to apply free probability techniques to numerically compute theasymptotic capacity of the channels under consideration. These theorems providea very efficient method for numerically approximating both the capacity and acapacity achieving input covariance matrix of certain channels.
展开▼
