Soft-Aided Staircase Code Decoding for Miscorrection Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing hard-decision-based FEC codes, such as staircase codes (SCCs) and product codes (PCs), suffer from limited error-correcting capability and miscorrections during decoding, which are not effectively addressed by current methods that either increase complexity or latency.
Innovation Solution
A soft-aided bit-marking (SABM) algorithm that marks highly reliable and highly unreliable bits to improve miscorrection-detection and error-correcting capability, using log-likelihood ratios to determine bit reliability without requiring additional memory or tracking soft bits.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If soft-decision FEC codes are used to achieve large net coding gain, then error correction performance is improved, but power consumption and decoding latency increase significantly
Solution Approach 1:
The patent applies local quality by selectively using soft information only for marking highly unreliable bits rather than processing all bits with soft-decision decoding. The marking process identifies specific bit positions that benefit from soft information, while the actual correction uses hard-decision logic, achieving a localized application of soft-decision advantages without the global complexity overhead.
Solution Approach 2:
The patent implements partial action by using soft information partially - only to mark unreliable bits for potential correction, rather than fully utilizing soft information for all decoding operations. This partial use of soft information provides performance improvement over pure hard-decision while avoiding the full complexity and latency of soft-decision decoding.
2Device complexity
If bounded-distance decoding is used for simple implementation, then device complexity is reduced, but error-correcting capability is limited to t errors
Solution Approach 1:
The patent introduces an intermediary marking process that sits between the received signal and the bounded-distance decoding. The marking step identifies unreliable bits using soft information, and this marked information guides the subsequent hard-decision decoding, effectively mediating between the simple BDD approach and the need for enhanced error correction capability.
Solution Approach 2:
The patent applies preliminary action by performing the marking of unreliable bits before the actual decoding process. This preliminary identification of problematic bit positions allows the subsequent hard-decision decoding to focus computational resources on correcting only the marked bits, thereby extending error correction capability beyond the standard BDD limit while maintaining algorithmic simplicity.
3Reliability
If channel soft information is used to extend error correcting capability beyond t errors, then reliability is improved, but decoding complexity increases due to multiple test pattern decodings
Solution Approach 1:
The patent extracts only the essential information from soft channel data - specifically, the identification of highly unreliable bits - rather than using the full soft information for multiple test pattern decodings. This extraction approach captures the most critical aspect of soft information for error correction while avoiding the complexity of exhaustive test pattern evaluation.
Solution Approach 2:
The patent uses soft information partially - only to mark unreliable bits for correction - rather than exhaustively decoding multiple test patterns as in traditional soft-decision methods. This partial utilization of soft information achieves extended error correction capability while maintaining complexity comparable to hard-decision decoding.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A hard-decision (HD) forward error correcting (FEC) coded signal is decoded by a soft-aided decoder (112) to produce decoded bits (118) using marked reliable bits of the HD-FEC coded signal and marked unreliable bits of the HD-FEC coded signal (soft-aided bit marking) that are computed by calculation (114) and marking blocks (116) based on an absolute value of log-likelihood ratios (LLRs) of the HD-FEC coded signal. The hard-decision (HD) forward error correcting (FEC) coded signal may be, for example, a staircase code (SCC) coded signal or a product code (PC) coded signal.