The vector perturbation precoding for massive multiple-input multiple-output based on GMD

In this paper, we think about the massive multiple-input multiple-output system analysis of vector perturbation precoding based on channel matrix geometric mean decomposition employing the minimum mean square error criterion with instantaneous power constraints for broadcast channels. Our analysis assumes that the transmitter has the imperfect channel state information and the receivers are acquainted with a perfect channel-dependent transmit power normalization factor. We propose a descent method by iteration for searching the perturbation vector in the complex field. The performances of the devised precoding scheme with the large quantity of transmit antennas and single-antenna users are studied in respect to bit error rate analysis through simulations and the numerical results illustrate it appropriated for massive multiple-input multiple-output system both in terms of complexity and performance.