Capacity and Complexity Trade-offs in MIMO Analog–Digital Combining Systems

Abstract

Analog–digital combining is an alternative solution to decrease the hardware complexity of an all-digital combining system. We analyze the analog weights from the perspectives of maximizing analog signal-to-noise ratio (SNR) and effective channel capacity ${(\tilde{C})}$. We propose simple sub-optimal algorithms to calculate receive analog weights in both open loop (OL) and closed loop (CL) systems, and compare the proposed schemes with an all-digital combining system, in terms of ${\tilde{C}}$. For both criteria, a small number of additional antennas is required to achieve the same mean capacity as an all-digital combining system in an urban micro model, e.g., 1 additional antenna in selection diversity and combining is required to achieve the same capacity as in a 2 ×  4 all-digital CL system at low SNR. The additional number of antennas is smaller in CL than in OL. The effective channel capacity based on the SNR maximization criterion is similar to that based on the ${\tilde{C}}$ maximization criterion, but the computational complexity the analog weights calculation is lower for the SNR maximization criterion.