Precoding Design for Distributed Antenna Systems in Spatially Correlated Ricean Fading Channel

Three precoding schemes that are based on statistical channel state information ( CSI) for distributed antenna systems ( DAS) in spatially correlated channels are presented. The first two schemes exploit the channel correlation at the transmitter and the eigenstructure of the precoder to develop efficient algorithms to obtain its beamforming matrix and power allocation ( PA). Based on the analysis of an average pairwise error probability bound at a low signal-to-noise ratio ( SNR), the first scheme derives a suboptimal unitary matrix that helps develop an efficient modified power adaptation. The scheme has low computational complexity but encounters slight performance loss in moderate- and high-SNR regions. To reduce the performance loss, the unitary matrix is reoptimized by using the Riemannian steepest descent method in the second scheme. With the unitary matrix and modified PA, a suboptimal precoder is obtained. Using the first precoding scheme, a low-complexity precoder design is developed for Rayleigh channels, which has closed-form power control in the form of a waterfilling rule. Compared with the existing precoding scheme that relies on a fixed-point numerical search, the proposed schemes have lower complexity. Simulation results show that the proposed schemes can provide symbol error performance better than the equal-PA scheme and close to the existing precoding scheme.