Cooperative Relaying Based on Alamouti Diversity under Aggregate Relay Power Constraints

This paper investigates power allocation in cooperative relaying based on Alamouti diversity. The objective is to find channel capacity optimal ways to allocate power between two relays under aggregate relay power constraints. Both a nonregenerative and a regenerative case are evaluated. The analysis shows that the total power should be allocated to one of the two relays, provided the instantaneous channel state information of both the 1^st hop and the 2^nd hop are available at the relays, i.e. for both non-regenerative and regenerative relaying. For the scenario that the relays have the instantaneous channel state information of the 1^st hop, but only the average channel state information of the 2^nd hop, the efforts focus on the regenerative relays while assuming the relays could perform decoding perfectly. Several optimization schemes: maximization of the mean channel capacity, maximization of the mean signal-to-noise ratio (SNR) and minimization of SNR variance are investigated. The "minimization of the SNR variance" scheme results in significantly lower mean channel capacity than the other two schemes do. The "maximization of the mean SNR" scheme still leads to the conclusion that the total power should be allocated to one of the two relays. The "maximization of the mean channel capacity" scheme gives almost the same mean channel capacity compared to the "maximization of the mean SNR" scheme and the two schemes are equivalent to each other in most cases