Bayesian Robust Linear Transceiver Design for Dual-Hop Amplify-and-Forward MIMO Relay Systems

Abstract

In this paper, we address the robust linear transceiver design for dual-hop amplify-and-forward (AF) MIMO relay systems, where both transmitters and receivers have imperfect channel state information (CSI). With the statistics of channel estimation errors in the two hops being Gaussian, we formulate the robust linear-minimum-mean-square-error (LMMSE) transceiver design problem using the Bayesian framework, and derive a closed-form solution. Simulation results show that the proposed algorithm reduces the sensitivity of the relay system to channel estimation errors, and performs better than the algorithm using estimated channel only.

Robust Design for Linear Non-Regenerative MIMO Relays With Imperfect Channel State Information

“…With mismatch between the true and the estimated CSI, the true channel H i can be represented by the well-known Gaussian-Kronecker model [5]- [9], where H i is a complexvalued Gaussian random matrix…”

“…In the first example, the true channel matrices are modelled as (3) with Θ i = I N i , i = 1, 2, and Φ i , i = 1, 2, are defined as Toeplitz matrices [5]- [9] with elements given by [Φ 1 ] m,n = 0.8 |m−n| and [Φ 2 ] m,n = 0.9 |m−n| , respectively. Fig.…”

“…In particular, we consider two imperfect CSI scenarios: (1) All nodes have imperfect CSI; (2) The destination node knows the exact CSI, while the source and relay nodes have imperfect CSI. The true CSI is modelled as a Gaussian random matrix with the estimated CSI as the mean value, and the well-known Kronecker model is adopted for the covariance of the CSI mismatch [5]- [9]. We would like to point out that in [10], only the imperfect CSI case (1) is addressed.…”