Multiple-input multiple-output (MIMO) wireless systems use antenna arrays at both the transmitter and receiver to provide communi- cation links with substantial diversity and capacity. Spatial multiplexing is a common spaceŠtime modulation technique for MIMO communication systems where independent information streams are sent over different transmit antennas. Unfortunately, spatial multiplexing is sensitive to ill- conditioning of the channel matrix. Precoding can improve the resilience of spatial multiplexing at the expense of full channel knowledge at the transmitterŃwhich is often not realistic. This correspondence proposes a quantized precoding system where the optimal precoder is chosen from a finite codebook known to both receiver and transmitter. The index of the optimal precoder is conveyed from the receiver to the transmitter over a low-delay feedback link. Criteria are presented for selecting the optimal precoding matrix based on the error rate and mutual information for different receiver designs. Codebook design criteria are proposed for each selection criterion by minimizing a bound on the average distortion assuming a Rayleigh-fading matrix channel. The design criteria are shown to be equivalent to packing subspaces in the Grassmann manifold using the projection two-norm and FubiniŠStudy distances. Simulation results show that the proposed system outperforms antenna subset selection and performs close to optimal unitary precoding with a minimal amount of feedback.
Diversity methods, Grassmannian subspace packing, MIMO systems, Quantized precoding, Rayleigh channels, Spatial multiplexing, V-BLAST.
This paper is available as a .IEEE Xplore .