Performance of unitary space-time modulation in a continuously changing channel

The fading channel is a significant problem in many communications environments. We examine the performance of unitary space-time modulation in a time-varying channel. We use a Gauss-Markov model of the continuously varying channel to characterize performance of differential and trained modulation. We find a performance ceiling at high SNR where the effect of the changing channel dominates. We show that while trained modulation provides an advantage at low SNR, above a certain SNR differential modulation gives better performance. We conclude with simulation results that support our analysis