Decentralized and Centralized IDD Schemes for Cell-Free Networks

TL;DR

This work tackles interference in cell-free mMIMO with APs selection by developing iterative interference cancellation schemes and LDPC-based iterative detection and decoding. It provides centralized and decentralized IDD architectures, deriving closed-form MMSE-Soft-IC detectors that account for imperfect CSI and APs selection, plus a List-MMSE-Soft-IC detector and three LLR-processing strategies. Through BER simulations, it demonstrates that decentralized processing with LLR refinement approaches centralized performance while reducing fronthaul signaling and computational load. The findings offer scalable, practical options for uplink CF-mMIMO systems using LDPC codes and APs selection.

Abstract

In this paper, we propose iterative interference cancellation schemes with access points selection (APs-Sel) for cell-free massive multiple-input multiple-output (CF-mMIMO) systems. Closed-form expressions for centralized and decentralized linear minimum mean square error (LMMSE) receive filters with APs-Sel are derived assuming imperfect channel state information (CSI). Furthermore, we develop a list-based detector based on LMMSE receive filters that exploits interference cancellation and the constellation points. A message-passing-based iterative detection and decoding (IDD) scheme that employs low-density parity-check (LDPC) codes is then developed. Moreover, log-likelihood ratio (LLR) refinement strategies based on censoring and a linear combination of local LLRs are proposed to improve the network performance. We compare the cases with centralized and decentralized processing in terms of bit error rate (BER) performance, complexity, and signaling under perfect CSI (PCSI) and imperfect CSI (ICSI) and verify the superiority of the distributed architecture with LLR refinements.

Figures (6)

• Figure 1: BER versus SNR while comparing detectors for decentralized processing for $L=4$, $N=4$, $K=4$: (a) Before LLR Refinement and (b) After LLR Refinement.
• Figure 2: BER versus SNR for All APs comparing decentralized and centralized processing for the case with imperfect CSI with $L=4$, $K=4$, $N=4$, $\mathrm{IDD}=2$.
• Figure 3: BER versus SNR for All APs comparing LLR Censoring and LLR Refinement for decentralized processing for the case with imperfect CSI with $L=4$, $K=4$, $N=4$, $\mathrm{IDD}=2$.
• Figure 4: BER versus SNR for a case that uses All APs and a case that uses APs-Sel for $L=4$, $N=4$, $K=4$: (a) Centralized Processing and (b) Decentralized Processing
• Figure 5: BER versus SNR while varying number of IDD iterations for $L=4$, $N=4$, $K=4$: (a) SIC and (b) List-SIC.