FDD based Cell-free massive MIMO with angular reciprocity over rician fading channels

We present a combined angle reciprocity and Rician K factor based feedback for up/downlink scenarios using MMSE estimator for frequency division duplex (FDD) based massive Multiple Input Multiple Output (MIMO) system. It considers phase variation for spatially correlated Rician fading channel and also takes into account the phase shift in the line-of-sight component of Rician channel that is incurred due to user mobility and/or phase noise. The cell-free massive MIMO system has largely been implemented in the context of time division duplex (TDD) systems. However, FDD is still prevalent in communication standards and therefore, it is essential to assess its performance for FDD based channel estimation schemes in a typical cell-free massive MIMO environment. To begin with, we present an analysis of estimation schemes including the minimum mean square error (MMSE), linear MMSE (LMMSE), and least-square (LS) in terms of their spectral efficiency, using prior channel knowledge at the access points. We also present an angle-based beamforming technique that combines the quantized channel feedback based on Rician K factor with angle reciprocity-based channel estimates. The multipath channel estimation for MMSE and LMMSE schemes is also presented and simulation results are provided that show the superior performance of proposed angle and feedback MMSE over the LMMSE and theoretical MMSE schemes. Numerical simulations of the overall system confirm that the combined angular reciprocity and feedback-based MMSE estimates perform better with a value of 12.6% for spectral efficiency and 27% for energy efficiency, thus, producing an overall improved system performance compared to the conventional FDD based estimation schemes.