Abstract
The coding gain and coding advantage criteria commonly used to analyse space-time codes are revisited. These criteria are based on the approximation of the average pairwise error probability achievable over a Rician multiple antenna channel at high signal-to-noise ratios. Both criteria appear to be inaccurate for a moderate number of antennas, especially in the environments with limited diversity and Rayleigh fading. In these cases, the average error probability may differ substantially from the outage probability for practically considered outage rates. A closed form upper bound on the outage pairwise error probability is derived. This bound gives rise to an alternative measure of coding advantage and coding gain that depend on the outage rate.

This publication has 1 reference indexed in Scilit: