Radio Frequency Distortion Compensation Scheme of Massive MIMO System Transmitter Based on Multi-channel Real-Valued Time-delay Neural Network

Massive MIMO (Multiple-Input Multiple-Output) system has the advantages of high spectrum efficiency and large system capacity. However, due to the limitation of technique and manufacture in transmitter's radio frequency components, massive MIMO systems inevitably have many serious radio frequency distortion problems, which have become an important bottleneck restricting the system performance. This article focuses on the radio frequency distortion problems of power amplifier (PA) nonlinearity, in-phase/quadrature (I/Q) branch imbalance, and RF link crosstalk in massive MIMO systems. The RF distortion compensation scheme of massive MIMO system transmitter based on Multi-channel Real-Valued Timed-Delay Neural Network (RVTDNN) is proposed. This scheme uses multiple RVTDNN predistortion networks to pre-process the transmitted signal, compensate for the undesirable radio frequency characteristics of the transmitter, and improve system performance. In addition, this paper also proposes a predistortion network hyperparameter optimization scheme based on quantum genetic algorithm. Compared with the traditional genetic algorithm-based network hyperparameter optimization scheme, this scheme can obtain the optimal solution with small population number and low time complexity.