Efficient linearisation technique for crosstalk and power amplifier non-linearity suitable for massive MIMO transmitters

Massive multi-input-multi-output (MIMO) is expected to be an eminent technique to meet the demands of high system capacity and data rates of wireless networks in 5G wireless communication. However due to inherent imperfections of the transmitter such as power-amplifier (PA) non-linearity and crosstalk, practically, the signal quality suffers and does not reap sufficient benefits from the various MIMO techniques. Digital predistortion (DPD) is a popular technique for single-input-single-output transmission to enhance signal quality. This study examines the issue of high DPD's complexity in mitigating imperfections in MIMO transmitters. This work proposes a less complex, novel DPD model for linearising large-scale MIMO transmitters along with its characterisation procedure. The proof-of-concept is provided with the measurement setup containing 4x1 MIMO transmitters in the presence of non-linear crosstalk, linear crosstalk, and strong PA non-linearity. The proposed model performs comparably to the state-of-art DPD models like parallel Hammerstein and dual-input crosstalk mismatch with lower number of floating-point operations (flops). The proposed model improves adjacent channel power ratio up to -54.13dBc and error vector magnitude up to 1.08% for LTE signal of 40MHz bandwidth.