Impact of Multiple Primaries and Partial CSI on Transmit Antenna Selection for Interference-Outage Constrained Underlay CR

Abstract

Transmit antenna selection is a low-complexity multiple-antenna technique that exploits spatial diversity using only one radio frequency chain. We investigate it for an underlay cognitive radio system that operates in the presence of multiple primary receivers and is subject to a constraint on the interference outage it causes at any of the primary receivers. The selection is based on a practically motivated and general partial channel state information (CSI) model in which the secondary transmitter (STx) only knows the channel power gains to a subset of the primary receivers. We derive a novel and general antenna selection rule that provably minimizes the symbol error probability (SEP) of the secondary system. We also derive insightful analytical expressions for its average SEP and interference-outage probability. These apply to a general class of channel fading models and any number of transmit and receive antennas, and include the special cases in which the STx knows channel power gains of all or none of the primary receivers. Our numerical results bring out a new insensitivity of the average SEP of the optimal rule to the interference power threshold when the CSI available is partial.