TL;DR
This paper introduces new methods for designing and analyzing "semi-structured" and "protograph" E²RC codes, a type of rate-compatible LDPC code. By developing a fast EXIT chart computation method and novel design rules, they achieve codes with performance very close to the theoretical capacity limit across a wide range of code rates, while also addressing issues like high error floors found in previous E²RC codes.
What Problem Does This Paper Address?
Rate-compatible punctured LDPC (RC-LDPC) codes are highly desirable because a single encoder/decoder pair can operate at multiple code rates. This is achieved by puncturing (removing) parity bits from a "mother code" designed for the lowest rate. However, designing good RC-LDPC codes that maintain high performance across all rates, especially when puncturing, is challenging. Previous E²RC codes offered efficient encoding and good puncturing performance but suffered from high error floors at the mother code rate and lacked systematic design methods for their structured parity parts. This paper aims to address these limitations by proposing new code structures and design techniques.
Methods
The authors introduce two main types of E²RC codes:
- Semi-structured E²RC-like codes: These codes have a deterministic parity part (H₂) with a protograph structure and an unstructured systematic part (H₁). They develop EXIT chart-based methods to find near-optimal degree distributions for H₁, taking into account the structure of H₂. A key innovation is a fast, analytical method for computing EXIT functions of code components with protograph structures, avoiding slow Monte Carlo simulations. This method allows for the joint optimization of code performance across multiple rates.
- Protograph E²RC codes: These codes extend the protograph structure to both the systematic (H₁) and parity (H₂) parts. They propose design rules for constructing families of rate-compatible punctured protographs with low thresholds. This structured approach is beneficial for practical implementation due to easier parallelization and reduced storage.
Datasets
Not discussed in the paper. The paper focuses on theoretical design and simulation-based performance evaluation of constructed codes.
Key Insights and Ideas
- Fast EXIT Function Computation: The development of a non-Monte Carlo method for computing EXIT functions of protograph-based code components is a significant contribution. This dramatically speeds up the design and analysis process.
- Joint Optimization Across Rates: The paper demonstrates a systematic way to design codes that maintain a small gap to capacity not just at a single rate, but across an entire range of puncturing rates. This is crucial for practical applications where the required rate can vary.
- Alleviating Error Floors: By imposing a protograph structure on the parity part (H₂) of semi-structured codes, the authors effectively reduce the maximum check node degree, which is a known cause of high error floors in earlier E²RC codes.
- Protograph Structure Benefits: The protograph representation for both H₁ and H₂ offers a structured way to design codes that are easier to implement in hardware and can be systematically expanded to achieve rate compatibility.
Key Results
- Near Capacity Performance: Both semi-structured and protograph E²RC families achieve gaps to capacity of at most 0.3 dB across a range of rates, with a maximum variable node degree of twenty.
- Improved Error Floor Performance: The proposed codes exhibit significantly reduced error floors compared to original E²RC codes, even outperforming some existing methods designed to address this issue.
- Faster Design Process: The new EXIT function computation method reduces the time for obtaining code performance predictions from days (Monte Carlo) to seconds.
- Competitive Performance: Protograph E²RC codes show performance comparable to or better than optimized semi-structured E²RC codes and existing literature codes (e.g., AR4JA family) at various rates.
Limitations and Future Work
- Limitations:
- The paper primarily focuses on the Binary-Input AWGN (BIAWGN) channel, though it notes the methods can be adapted to other channels.
- While the protograph construction offers advantages, the search for optimal splitting patterns can still be computationally intensive, though significantly reduced by the proposed heuristics.
- Future Work:
- The authors suggest their methods are applicable to other structured code components like IRA codes.
- The paper lays the groundwork for further exploration of protograph-based RC-LDPC codes and their efficient implementation.
- The joint optimization algorithm could be further refined for specific application requirements.
Summary updated