Highly Nonlinear and Wide-Band mmWave Active Array OTA Linearization Using Neural Network

This paper proposes a neural network (NN) based over-the-air (OTA) linearization technique for a highly nonlinear and wide-band mmWave active phased array (APA) transmitter and compares it with the conventional memory polynomial model (MPM)basedtechnique.TheproposedNNeffectivelylearnsthedistinctivenonlineardistortions,which may not easily fit to existing MPM solutions, and can therefore successfully cope with the challenges introduced by the high nonlinearity and wide bandwidth. The proposed technique has been evaluated using a state-of-the-art 4x4 APA operating in highly nonlinear regions at 28 GHz with a 100 MHz wide 3GPP base-station signal as input. Experimental results show the predistortion signal generated by the NN exhibits peak to average power ratio (PAPR) much lower than the one generated by MPM and consequently superior linearization performance in terms of adjacent channel leakage ratio (ACLR) and error vector magnitude (EVM) for high nonlinearity cases. Using the proposed NN based linearization technique, an improvement of 5 dB ACLR and 7 % points in EVM are achieved, which demonstrates the promising potential of this technique for emerging broadband communication systems such as 5G/6G and low earth orbit (LEO) satellite networks.