Auxiliary Variable-Aided Hybrid Message Passing for Joint Channel, Phase Noise Estimation and Detection for Multi-User MIMO Systems

This paper considers a multi-user multiple-input multiple-output (MU-MIMO) system with independent phase noise (PN) at user terminals (UTs) due to non-synchronous noisy local oscillators. The conventional factor graph of MU-MIMO involves an observation multiple high-dimensional variables in the form multiplication and summation. In this work, is represented by introducing auxiliary variables, which enables decomposition into several simpler sub-factors then use different message passing techniques obtain approximate marginals. Specifically, belief propagation expectation (BP-EP) are used for linear factors random walk process, modulation coding, while mean field (MF) non-linear exponential form. To reduce computational complexity, non-Gaussian MF messages approximated be Gaussian using second order Taylor expansion its obtained previous iteration. proposed receiver has quadratic complexity per symbol Simulation results show that performance better than existing methods close matched filter bound (MFB) terms bit error rate (BER).