Parallel Channel Bidiagonalization for Massive MIMO Systems

Recently, the channel bidiagonalization (CBD) [1] method is proposed to decompose the multiple-input multiple-output (MIMO) channel into a bidiagonal channel, so that the maximum likelihood (ML) or maximum a posteriori (MAP) detection and decoding can be adopted in a low complexity. Inspired by the advantages of the bidiagonal structure, in this paper, we propose a novel linear transceiver design, termed as parallel CBD (PL-CBD), which supports parallel optimal detection at the receiver with a lower complexity than the original CBD, thereby providing a practical solution for massive MIMO system. Moreover, the PL-CBD exhibits improved bit error rate (BER) performance compared to the original CBD. The simulation results in a 5G New Radio (5G-NR) simulator also validate the superiority of PL-CBD over other state-of-the-art algorithms.