Channel Estimation for Spatial Modulation Schemes in Spatially Correlated Time Varying Channels

In this paper, we propose a channel estimation scheme for spatial modulation (SM) systems. In general, SM systems require each transmit antenna to separately send pilot symbols for channel estimation. This is a lengthy process, which may incur significant channel estimation errors in a time varying channel. Thus, we propose correlation-based channel estimation (CBCE) scheme, which exploits the correlation between transmit antennas to estimate channels of inactive antennas using the pilot-based estimate of the active antenna. The change in the active channel from the last pilot-based estimate is calculated, and a time-proportionate amount of that change is scaled according to the channel correlation coefficients to estimate the channel state of the inactive channels. With pilot slots designed to spread out in a data frame, the estimation process for every channel is carried out N_t times in a data frame with N_t denoting the number of transmit antennas. We observe that in a high signal-to-noise ratio (SNR) regime, our proposed scheme provides about 2 dB and 5 dB gains compared to conventional channel estimation (CCE) method for moderately correlated and highly correlated antennas, respectively. Through Monte Carlo simulations with different correlation ρ and user speeds, we validate our analysis and show that the proposed scheme outperforms CCE when ρ ≥ 0.3, while it provides comparable performance for small ρ.