Blind Joint Nonlinear Equalizer for mmWave Communications

Abstract

In this paper, we study the blind joint nonlinear receiver for millimeter wave (mmWave) communication systems. The mmWave system operates in a very high-frequency band and has a lot of bandwidth resources. The RF power amplifier (PA) used by the system poses a huge challenge to signal detection and demodulation. Instead of using high-complexity and difficult-to-implement digital predistortion at the transmitter, the blind joint nonlinear receiver is proposed to make mmWave communication a new possibility in 5G NR. Due to the non-linearity and edge integration, the posteriori probability of the received signal is quite difficult to analyze, which makes existing linear equalization techniques unusable. Therefore, we use the particle filter that is based on Monte Carlo sequence importance sampling. With the aid of LDPC (low-density parity-check) decoding, the proposed blind receiver can operate without training sequence. In this way, the unknown transmitting signal can be successfully recursively estimated. The simulation results validate the proposed method.