Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Iterative Detection and Decoding for Clustered Cell-Free Massive MIMO Networks

T. Ssettumba, S. Mashdour, L. Landau, P. da Silva, R. C. de Lamare

TL;DR

This work tackles the performance degradation from intra-cluster and out-of-cluster interference in user-centric clustered CF-mMIMO uplinks. It proposes an integrated approach consisting of an MMSE receive filter with a modified PIC to cancel both ICL and OCL streams, a least-squares estimator for OCL interference, and an LDPC-based iterative detection and decoding (IDD) chain with a Gaussian-approximation receiver. Key contributions include a projection-based OCL estimation strategy leveraging $\tau_p > K+M$, a scalable modified PIC receiver, and a tractable IDD with box-plus SPA, achieving notable BER and NMSE gains over baselines. The results indicate practical improvements for interference management in scalable CF-mMIMO deployments, with implications for pilot design and receiver complexity.

Abstract

In this letter, we propose an iterative soft interference cancellation scheme for intra-cluster (ICL) and out-of-cluster (OCL) interference mitigation in user-centric clustered cell-free massive multiple-antenna networks. We propose a minimum mean-square error receive filter with a novel modified parallel interference cancellation scheme to mitigate ICL and OCL interference. Unlike prior work, we model the OCL interference and devise a least squares estimator to perform OCL interference estimation. An iterative detection and decoding scheme that adopts low-density parity check codes and incorporates the OCL interference estimate is developed. Simulations assess the proposed scheme against existing techniques in terms of bit error rate performance.

Iterative Detection and Decoding for Clustered Cell-Free Massive MIMO Networks

TL;DR

This work tackles the performance degradation from intra-cluster and out-of-cluster interference in user-centric clustered CF-mMIMO uplinks. It proposes an integrated approach consisting of an MMSE receive filter with a modified PIC to cancel both ICL and OCL streams, a least-squares estimator for OCL interference, and an LDPC-based iterative detection and decoding (IDD) chain with a Gaussian-approximation receiver. Key contributions include a projection-based OCL estimation strategy leveraging , a scalable modified PIC receiver, and a tractable IDD with box-plus SPA, achieving notable BER and NMSE gains over baselines. The results indicate practical improvements for interference management in scalable CF-mMIMO deployments, with implications for pilot design and receiver complexity.

Abstract

In this letter, we propose an iterative soft interference cancellation scheme for intra-cluster (ICL) and out-of-cluster (OCL) interference mitigation in user-centric clustered cell-free massive multiple-antenna networks. We propose a minimum mean-square error receive filter with a novel modified parallel interference cancellation scheme to mitigate ICL and OCL interference. Unlike prior work, we model the OCL interference and devise a least squares estimator to perform OCL interference estimation. An iterative detection and decoding scheme that adopts low-density parity check codes and incorporates the OCL interference estimate is developed. Simulations assess the proposed scheme against existing techniques in terms of bit error rate performance.

Paper Structure

This paper contains 10 sections, 30 equations, 4 figures, 1 table.

Table of Contents

  1. Introduction
  2. Proposed System Model, Channel and Interference Estimation
  3. Channel Estimation for the Serving Users
  4. Interference Estimation
  5. Uplink Data Transmission
  6. Proposed Iterative Detection and Decoding
  7. Proposed Iterative Receiver Design
  8. IDD Scheme with Gaussian Approximation
  9. Simulation Results
  10. Concluding Remarks

Figures (4)

  • Figure 2: NMSE versus SNR $L=32$, $N=4$, $K=8$, $M=4$ for OCL interference estimation and channel estimation.
  • Figure 3: BER versus SNR $L=32$, $N=4$, $K=8$, $M=4$ for the system with ICL and OCL interference, QPSK, and a varying number of modified PIC IDD iterations.
  • Figure 4: BER versus SNR for (a) $L=32$, $N=4$, $K=8$, $M=4$, QPSK, (b) $L=32$, $N=4$, $K=8$, $M=2$, QPSK and (c) $L=32$, $N=4$, $K=8$, $M=2$, 16-QAM for the clustered network.
  • Figure : (a)