Hybrid wideband transceiver design for achievable rate maximization in millimeter-wave MIMO systems

This paper proposes two algorithms for hybrid (Analog-Digital) beamforming in a single-user millimeter-wave (mm-wave) multi-input multi-output (MIMO) systems under frequency selective channels. Hybrid architecture has a potential to reduce the number of radio frequency (RF) chains to far less than the number of antennas and, it results in low hardware complexity, power efficiency and cost-saving to the system. Since practical broadband systems are frequency selective in nature therefore, the challenging aspect of this problem is to have only one analog RF processing matrix at source and destination over the entire frequency band. To address this problem for spectral efficiency maximization, the first algorithm performs multi-linear singular value decomposition on different sub-carriers to derive the common analog combiner over the allocated frequency-spectrum. Then, a problem is formulated for designing the common analog precoder that attempts to extract the optimal phase values for gain maximization of the equivalent channel observed from baseband processing units at transmitter and receiver. To significantly reduce the computational complexity in comparison to the first algorithm, the second algorithm employs classification of beamforming vectors for gain maximization of the equivalent baseband channel over all sub-carriers to find the analog RF precoding/combining matrices. Digital baseband processing components are derived by diagonalizing the equivalent baseband channel corresponding to each frequency-carrier. Computational complexity of the second algorithm is significantly less than the several existing designs. Computer simulations reveal that the proposed algorithms show near-optimal performance under diverse scenarios of system configuration parameters.