Search

In this paper, an emerging wireless communication concept, which is termed as spatial modulation (SM) for large-scale multiple-input–multiple-output (MIMO), is considered. The results show that from the information-theoretic perspective, SM achieves capacity comparable to the open-loop MIMO capacity, although a subset of transmit antennas is activated in every channel use because both the channel coefficients and the input symbols carry information in SM. As a result, SM compensates the loss of information capacity due to a subset of antennas being activated by modulating information in the antenna index; therefore, the total information rate remains high. In particular, an upper bound for the capacity of SM is derived in closed form, and it is shown analytically that this upper bound is almost certainly achievable in the massive MIMO regime. Moreover, it is shown that the upper bound is achievable with no channel state information at the transmitter (CSIT) but with channel distribution information (CDI) at the transmitter (CDIT). The optimum transmission strategy should adapt the channel input distribution to fading using CDI such as the $K$ factor in Rician fading or the shape parameter $m$ in Nakagami- $m$ fading. [ABSTRACT FROM PUBLISHER]

The prospect of base station cooperation leading to joint combining at widely separated antennas has led to increased interest in macrodiversity systems, where both sources and receive antennas are geographically distributed. In this scenario, analytical investigation of channel capacity is extremely challenging for finite-size systems since the channel matrices have a very general form where each path may have a different power. Hence, in this paper, we consider the ergodic sum capacity of a macrodiversity multiple-input multiple-output system with arbitrary numbers of sources and receive antennas operating over Rayleigh fading channels. For this system, we compute the exact ergodic capacity for a system with at most two transmit antennas and a compact approximation for the general system, which is shown to be very accurate over a wide range of cases. Finally, we compare our results with previous asymptotic results and bounds. Results are verified by Monte Carlo simulations and the impact on capacity of various channel power profiles is investigated. [ABSTRACT FROM AUTHOR]