Optimal predictive TX combining diversity in correlated fading for microcellular mobile radio applications

Transmitter diversity uses several transmit antennas to improve the performance of digital transmission in a fading environment. In transmitter selection diversity the best transmit antenna is selected on the basis of estimates of the propagation conditions from each antenna to the single receive antenna. In time division duplex systems these estimates can be obtained by measuring the received signal at each antenna in the receive time slots, which precede the transmit time slot. It has been proposed to use a predictor to improve the performance for a moving receiver (predictive TX selection diversity). No analysis is available in correlated fading. We consider predictive transmitter combining diversity in contrast to selection diversity this scheme uses all transmit antennas simultaneously. Our main contribution is a derivation and an analysis of the optimal predictive TX combining diversity algorithm in correlated fading. We derive exact closed form expressions for the average probability of error for two suboptimal combining methods and tight bounds on the performance of the optimal combining method. Quantitative results on the average bit error rate and the diversity gain are given for a system with DECT parameters