Diversity Analysis of Randomized Linear Dispersion Codes in a Half-Duplex Amplify-and-Forward Multiple-Relay System

Abstract

A point-to-point multiple-relay communication system with half-duplex constraints is considered. The relays operate under the amplify-and-forward paradigm and implement a linear dispersion distributed space-time code with randomized dispersion matrices of independent and identically distributed entries. The large-signal-to-noise-ratio behavior of the asymptotically large deterministic system is studied for both the optimum (maximum likelihood) receiver and the linear minimum mean squared error (LMMSE) receiver. The diversity order of the system is shown to have a strong dependence on the aspect ratio of the linear dispersion matrices. More specifically, when the linear dispersion matrices are sufficiently tall, both receivers achieve the maximum diversity order, i.e., the total number of relays plus one. Conversely, when the linear dispersion matrices are fat (in the sense that relays linearly compress the information received from the source), the optimum receiver is shown to achieve an asymptotic diversity order of two, whereas the LMMSE receiver is totally unable to exploit the available spatial diversity.