Low Complexity Codebook-Based Beamforming for MIMO-OFDM Systems in Millimeter-Wave WPAN

In this paper, we consider a beamforming system in millimeter-wave wireless personal area networks (WPAN), where transmit and receive weight vectors are jointly derived by exchanging a training sequence repeatedly with different combinations of transmit and receive weight vectors. We propose a low complexity codebook-based beamforming scheme that consists of multiple levels and level-adaptive antenna selection in order to reduce the beamforming setup time. For each level, 1) the transmit and receive antennas are selected according to pre-determined inter-element spacings, 2) the training sequences are sent with different weight vectors from a pre-defined codebook and 3) the receiver selects the best transmit and receive weight vectors in order to optimize an effective signal-to-noise ratio (SNR) and these vectors are used to determine the codebook for the next following level. Even with low complexity, we show from the numerical results that our proposed scheme can provide an effective SNR gain and an average spectral efficiency approaching those of the codebook-based beamforming with exhaustive search.