Statistical Channel State Information Aided Local Precoding Optimization for Cell Free Massive MIMO Systems

Cell-free (CF) massive multiple-input multipleoutput (MIMO) system based on local operations is recognized as a viable next-generation network architecture with high performance. Leveraging large-dimensional random matrix theory, we have derived a deterministic equivalent performance expression for the performance of a CF MIMO system employing local regularized zero-forcing (L-RZF) precoding, and based on this performance expression, we have designed a high-performance regularization parameter optimization method, termed statistical channel state information aided L-RZF (SCSI-L-RZF). Under SCSI-L-RZF method, the CPU only needs to optimize the regularization parameters for each AP by utilizing statistical channel state information (CSI), while each AP only needs to design precoding by combining the regularization parameter and local CSI to achieve good performance gains. Simulation results demonstrate that our SCSI-L-RZF scheme outperforms the current mainstream maximum ratio transmission (MRT) precoding schemes, L-RZF and local team minimum mean square error (LT-MMSE), further narrowing the performance gap between local and global precoding.