[Paper Review] A Low-decoding-complexity, Large coding Gain, Full-rate, Full-diversity STBC for 4 X 2 MIMO System

K. Pavan Srinath, B. Sundar Rajan|ArXiv.org|Jun 11, 2008

Advanced Wireless Communication Techniques11 references3 citations

TL;DR

This paper proposes a novel full-rate, full-diversity space-time block code (STBC) for 4×2 MIMO systems that achieves low decoding complexity (order M⁵) while offering superior coding gain and error performance compared to existing codes like DjABBA and BHV. The code leverages coordinate interleaved orthogonal designs (CIODs) and enables efficient sphere decoding via structured decomposition, outperforming prior methods in both diversity gain and error rate for QAM constellations.

ABSTRACT

This paper proposes a low decoding complexity, full-diversity and full-rate space-time block code (STBC) for 4 transmit and 2 receive ($4 imes 2$) multiple-input multiple-output (MIMO) systems. For such systems, the best code known is the DjABBA code and recently, Biglieri, Hong and Viterbo have proposed another STBC (BHV code) which has lower decoding complexity than DjABBA but does not have full-diversity like the DjABBA code. The code proposed in this paper has the same decoding complexity as the BHV code for square QAM constellations but has full-diversity as well. Compared to the best code in the DjABBA family of codes, our code has lower decoding complexity, a better coding gain and hence a better error performance as well. Simulation results confirming these are presented.

Motivation & Objective

• To design a full-rate, full-diversity STBC for 4×2 MIMO systems with reduced decoding complexity compared to existing codes.
• To overcome the limitations of the BHV code, which has lower decoding complexity than DjABBA but lacks full diversity.
• To improve upon the DjABBA code by achieving lower decoding complexity and better coding gain for QAM constellations.
• To enable efficient maximum-likelihood decoding through structured real-valued sphere decoding with reduced dimensionality.
• To provide a code that outperforms all known 4×2 STBCs in terms of error rate and diversity gain for QAM signaling.

Proposed method

• The proposed STBC is constructed using Coordinate Interleaved Orthogonal Designs (CIODs) to ensure full diversity and full rate.
• The code structure allows decomposition of the decoding process into an 8-dimensional real sphere decoding for symbols x₅ to x₈, followed by four parallel 2-dimensional real sphere decodings for x₁ to x₄.
• The decoding complexity is reduced to O(M⁵) for all complex constellations, and O(M⁵) for square QAM, significantly lower than DjABBA (O(M⁷)) and BHV (O(M⁶) for square QAM).
• The code uses a matrix decomposition approach where the channel matrix is transformed into a structured upper triangular matrix R, enabling efficient sequential decoding.
• The real-valued sphere decoding is applied to the transformed signal model, exploiting the block structure of R to decouple symbol groups.
• The method ensures that real and imaginary parts of each symbol are handled efficiently, with cross-coupling only within individual symbols, not across them.

Experimental results

Research questions

• RQ1 Can a full-rate, full-diversity STBC for 4×2 MIMO be designed with decoding complexity lower than that of the DjABBA code?
• RQ2 Does the proposed code achieve better coding gain and error performance than the BHV code, which has lower complexity but lacks full diversity?
• RQ3 Can the decoding complexity be reduced to O(M⁵) for all QAM constellations while maintaining full diversity and full rate?
• RQ4 Is it possible to achieve better codeword error rate (CER) performance than the best-known DjABBA code through improved coding gain?
• RQ5 Can the proposed code outperform existing codes in high-SNR regimes due to full diversity and enhanced coding gain?

Key findings

• The proposed STBC achieves a minimum determinant of 0.64 for 4-QAM, significantly higher than the DjABBA code (0.04) and the BHV code (0), indicating superior coding gain.
• The decoding complexity is O(M⁵) for all complex constellations, which is lower than the DjABBA code (O(M⁷)) and the BHV code (O(M⁶) for square QAM).
• For square QAM constellations, the proposed code has a worst-case decoding complexity of 4M⁵, compared to 2M⁷ for the BHV code, resulting in a substantial reduction.
• Simulation results show that the proposed code outperforms both the DjABBA and BHV codes at high SNR due to its full diversity and higher coding gain.
• In 16-QAM signaling, the proposed code maintains superior CER performance over the BHV code and slightly exceeds the DjABBA code at high SNR, confirming its improved error rate performance.
• The code is the best-performing known STBC for 4×2 MIMO systems with QAM constellations, combining low decoding complexity, full diversity, full rate, and superior error performance.