Soft LLR Correction Using Hard Decisions in Multilevel Modulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional soft decision value generating apparatuses face challenges in calculating log-likelihood ratios (LLRs) with high accuracy due to high computational requirements, especially for modulation schemes involving differential encoding and non-gray mapped QAM with odd power of 2 modulation levels.
Innovation Solution
A soft decision value generating apparatus comprising a hard decision value calculator, an approximate value calculator, and an approximate value corrector that inverts the sign bit of approximate LLRs when conflicting with hard decision values, allowing for the calculation of LLRs with high accuracy using reduced computational resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the exact LLR calculation using equation (1) is performed, then the accuracy of soft decision values is improved, but the computational complexity and circuit size become enormous
Solution Approach 1:
The patent transforms the exact LLR calculation (equation 1) into an approximate calculation (equation 2) by changing the computational parameters. Instead of computing all exponential functions and taking logarithms, the patent uses a simplified formula based on minimum distance calculations: Lk ≈ 4/(σ²) × min(|rk - sk,0,min|, |rk - sk,1,min|). This parameter change maintains reasonable accuracy while dramatically reducing computational complexity and circuit size requirements.
Solution Approach 2:
The patent employs a computationally inexpensive approximation method that sacrifices some precision for vastly reduced complexity. The approximate LLR calculation uses simple distance metrics and basic arithmetic operations instead of complex exponential and logarithmic functions, making the system practically implementable with standard hardware while achieving sufficient decoding performance.
2Device complexity
If the approximate LLR calculation using equation (2) is used, then the computational complexity is reduced, but the accuracy of soft decision values deteriorates
Solution Approach 1:
The patent incorporates a correction mechanism that uses hard decision values as feedback to refine the approximate LLR calculations. The hard decision values (obtained by comparing received symbols with constellation points) are used to verify and adjust the approximate LLR results, ensuring that the final soft decision values maintain high accuracy despite using the simplified calculation method.
Solution Approach 2:
The patent applies a two-stage approach where the approximate calculation provides initial LLR values, and then selective correction is applied based on hard decision results. This partial correction strategy focuses computational resources on improving accuracy only where needed, rather than performing full exact calculations for all symbols, thus achieving a balance between complexity and precision.
3Adaptability or versatility
If differential encoding is applied to multilevel modulation, then the adaptability of the communication system is improved, but the calculation of LLRs requires significantly more computation
Solution Approach 1:
The patent separates the LLR calculation into distinct components: the approximate calculation part (using equation 2) and the correction part (using hard decision values). For differentially encoded multilevel modulation, this segmentation allows the system to handle the complex differential decoding requirements while maintaining efficient LLR computation through the approximate-correct approach, rather than requiring a completely new complex calculation method.
Data Source
Figure 1~2
Figure 3
Figure 4~5
AI summary
There are provided a hard decision value calculator that calculates hard decision values from I-ch and Q-ch coordinates of a received symbol having been subjected to multilevel modulation, and an LLR calculator that approximately calculates LLRs from the I-ch and Q-ch coordinates of the received symbol. In a case where a sign bit of an LLR calculated by the LLR calculator conflicts with a hard decision value calculated by the hard decision value calculator, an LLR corrector inverts the sign bit of the LLR and outputs the sign-bit-inverted LLR as a soft decision value to an error correction decoder.