Dual Codebooks for MIMO Beamforming with No Direct Channel Estimation

In Multiple-input multiple-output (MIMO) wireless systems, beamforming (BF) at the transmitter, and beamcombining (BC) at the receiver are simple spatial diversity techniques that increase the effective signal to noise ratio (SNR) of the system. The optimal BF-BC systems require estimation of the channel at the receiver and transmission of the optimum beamforming vector back to the transmitter. In related research, authors define a codebook of beamforming vectors, and the receiver sends back the index of the vector that maximizes the SNR. In all of these methods channel estimation can be quite complex, time consuming and can lead to more errors especially when strong, unmodeled, interference is present. In this paper, a codebook method that does not require channel estimation at the receiver is proposed. There is a beamforming vector codebook at the transmitter and a beamcombining vector codebook at the receiver. The codebook vector selection algorithm is based on maximizing the received SNR. The losses are minimal when compared to the ideal BF-BC method and the method using a codebook only at the transmitter.