Massive multiple-input multiple-output (mMIMO) technology is considered a key enabler for the 5G and future wireless networks. In most wireless communication systems, mMIMO is employed together with orthogonal frequency-division multiplexing (OFDM) which exhibits a high peak-to-average-power ratio (PAPR). While passing the OFDM signal through one of the common RF front-ends of limited linearity, significant distortion of the transmitted signal can be expected. In mMIMO systems, this problem is still relevant as in some channels the distortion component is beamformed in the same directions as the desired signal. In this work, we propose a multi-antenna clipping noise cancellation (MCNC) algorithm for the downlink of the mMIMO OFDM system. Computer simulations show it can remove nonlinear distortion even under severe nonlinearity. Next, a simplified version of the algorithm is proposed. It was observed that for the direct visibility channels, its performance is only slightly degraded with respect to the MCNC algorithm.

翻译：大规模多输入多输出（mMIMO）技术被视为5G及未来无线网络的关键使能技术。在大多数无线通信系统中，mMIMO常与正交频分复用（OFDM）结合使用，而OFDM信号具有高峰均功率比（PAPR）。当OFDM信号通过线性度有限的常见射频前端时，发射信号可能产生显著失真。在mMIMO系统中，这一问题仍然存在，因为在某些信道中，失真分量会与期望信号沿相同方向进行波束赋形。本文针对mMIMO OFDM系统下行链路，提出了一种多天线削波噪声消除（MCNC）算法。计算机仿真表明，即使在严重非线性条件下，该算法也能消除非线性失真。随后，本文提出了该算法的简化版本。经观察，在直视信道条件下，简化算法的性能相较于MCNC算法仅略有下降。