A Low-Complexity Digital Predistorter for Large-Scale MIMO Systems With Crosstalk

Abstract

Massive multiple-input multiple-output (MIMO) technology is a key enabler for future wireless communication systems. On the other hand, crosstalk is inevitable in the highly integrated large-scale antenna array. The nonlinearity of power amplifiers (PAs) coupled with crosstalk is much more complicated when the number of antennas is large. Digital predistortion (DPD) can compensate for the PA nonlinearity without sacrificing efficiency and is widely applied in both base stations and mobile devices. Extending existing solutions to the large-scale MIMO system usually suffers from performance degradation or greatly increased complexity. In this paper, we consider the nonlinearity with crosstalk for the large-scale MIMO scenario. A much simplified DPD model is proposed by ignoring the high-order crosstalk terms generated from the complete nonlinear model. Compared to existing solutions, the proposed DPD can compensate for both the nonlinear and linear crosstalk of the system. It achieves satisfactory linearization performance with much reduced computational complexity. The complexity of the proposed DPD is linearly related to the number of transmitter branches. Thus, it is scalable in the large-scale MIMO scenario. Moreover, the proposed DPD can be extended to the partially-connected hybrid beamforming system with crosstalk. Simulation results validate the effectiveness of the proposed DPD scheme.