[Paper Review] Near Capacity Approaching for Large MIMO Systems by Non-Binary LDPC Codes with MMSE Detection

Puripong Suthisopapan, Kenta Kasai+2 more|arXiv (Cornell University)|Mar 5, 2012

Advanced Wireless Communication Techniques30 references3 citations

TL;DR

This paper proposes a non-binary LDPC-coded large MIMO system with low-complexity MMSE detection, achieving near-capacity performance. By leveraging soft-output MMSE detection and moderate-length non-binary LDPC codes over GF(2^8), the system achieves a bit error rate within 3.4 dB of MIMO capacity for a 600×600 MIMO system with 16-QAM, outperforming existing turbo-coded and binary LDPC-coded systems by over 6 dB in SNR gain.

ABSTRACT

In this paper, we have investigated the application of non-binary LDPC codes to spatial multiplexing MIMO systems with a large number of low power antennas. We demonstrate that such large MIMO systems incorporating with low-complexity MMSE detector and non-binary LDPC codes can achieve low probability of bit error at near MIMO capacity. The new proposed non-binary LDPC coded system also performs better than other coded large MIMO systems known in the present literature. For instance, non-binary LDPC coded BPSK-MIMO system with 600 transmit/receive antennas performs within 3.4 dB from the capacity while the best known turbo coded system operates about 9.4 dB away from the capacity. Based on the simulation results provided in this paper, the proposed non-binary LDPC coded large MIMO system is capable of supporting ultra high spectral efficiency at low bit error rate.

Motivation & Objective

• To close the performance gap between coded large MIMO systems and theoretical MIMO capacity.
• To investigate the feasibility and performance of non-binary LDPC (NBLDPC) codes in large MIMO systems with low-complexity detection.
• To evaluate the effectiveness of MMSE detection as a soft-input source for NBLDPC decoding in high-dimensional MIMO settings.
• To establish benchmark performance for future coded large MIMO system designs.

Proposed method

• Designing a non-binary LDPC code over GF(2^8) with a sparse parity-check matrix for improved error correction in high-order modulation.
• Implementing a low-complexity MMSE detector to provide soft-input to the NBLDPC decoder, avoiding the exponential complexity of optimal detection.
• Deriving soft-output LLR values from the MMSE detector using a log-likelihood ratio metric based on the conditional probability of received symbols.
• Integrating the MMSE soft-output with iterative decoding via message-passing on the NBLDPC code’s factor graph.
• Using iterative detection-decoding with a fixed number of iterations to balance performance and complexity.
• Simulating the system over Rayleigh fading channels with 16-QAM modulation for large MIMO configurations (e.g., 600×600).

Experimental results

Research questions

• RQ1 Can non-binary LDPC codes achieve near-capacity performance in large MIMO systems with low-complexity MMSE detection?
• RQ2 How does the performance of NBLDPC-coded large MIMO systems compare to existing turbo-coded and binary LDPC-coded systems in terms of bit error rate and SNR gap to capacity?
• RQ3 Why does the MMSE detector—despite its poor uncoded performance—yield excellent coded performance when paired with NBLDPC decoding?
• RQ4 What is the impact of soft-output generation quality from MMSE detection on the overall system performance in high-dimensional MIMO systems?
• RQ5 Can moderate-length NBLDPC codes achieve near-capacity performance in large MIMO systems without requiring optimal detection?

Key findings

• The proposed NBLDPC-coded large MIMO system achieves a bit error rate within 3.4 dB of the theoretical MIMO capacity for a 600×600 MIMO system with 16-QAM and code rate R=1/3.
• The system outperforms the best-known turbo-coded large MIMO system by more than 6 dB in SNR, which operates 9.4 dB from capacity under the same conditions.
• The NBLDPC-coded system with MMSE detection achieves better bit error rate performance than both binary LDPC-coded and turbo-coded systems across all simulated SNR regimes.
• The MMSE detector’s soft-output quality, despite its low diversity order, enables near-optimal performance when combined with NBLDPC decoding, particularly in the low-SNR operating region of coded systems.
• The performance gap to capacity is significantly reduced compared to prior works, establishing the NBLDPC-MMSE system as the best-performing coded large MIMO configuration reported in the literature.
• The results demonstrate that NBLDPC codes are a superior channel coding choice for large MIMO systems, especially when higher-order modulations are used.