On the Exact and Approximate Eigenvalue Distribution for Sum of Wishart Matrices

TL;DR

This paper derives exact and approximate eigenvalue distributions for sums of Wishart matrices, with applications to MIMO communication systems, providing closed-form formulas validated by simulations.

Contribution

It introduces a new closed-form expression for the eigenvalue distribution of weighted sums of Wishart matrices and an approximation method, advancing analysis in multiuser MIMO channels.

Findings

Exact eigenvalue distribution matches simulations perfectly.

The approximation closely matches the exact distribution with negligible difference.

Closed-form expressions for ergodic sum-rate capacity and upper bounds are provided.

Abstract

The sum of Wishart matrices has an important role in multiuser communication employing multiantenna elements, such as multiple-input multiple-output (MIMO) multiple access channel (MAC), MIMO Relay channel, and other multiuser channels where the mathematical model is best described using random matrices. In this paper, the distribution of linear combination of complex Wishart distributed matrices has been studied. We present a new closed form expression for the marginal distribution of the eigenvalues of a weighted sum of K complex central Wishart matrices having covariance matrices proportional to the identity matrix. The expression is general and allows for any set of linear coefficients. As an application example, we have used the marginal distribution expression to obtain the ergodic sum-rate capacity for the MIMO-MAC network, and the cut-set upper bound for the MIMO-Relay case, both as closed form expressions. We also present a very simple expression to approximate the sum of Wishart matrices by one equivalent Wishart matrix. All of our results are validated by means of Monte Carlo simulations. As expected, the agreement between the exact eigenvalue distribution and simulations is perfect, whereas for the approximate solution the difference is indistinguishable.