Soft FEC Candidate Winnowing for Lower Decoding Complexity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High data transmission networks face increased complexity and costs due to the doubling of candidate words generated by increasing the number of least reliable positions (LRPs) in soft iterative error correction techniques, leading to higher processing complexity and costs.
Innovation Solution
A system and method that identify least reliable positions in encoded words, generate candidate words based on these positions, and perform a winnowing operation to reduce the number of candidate words while maintaining error correction capabilities, thereby reducing processing complexity and costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the quantity of least reliable positions (LRPs) is increased to identify more errors, then the error correction capability is improved, but the quantity of candidate words doubles, increasing processing complexity and cost
Solution Approach 1:
The patent segments the set of least reliable positions into two groups: a first set of LRPs that are always processed, and a second set of LRPs that are conditionally processed based on whether errors were detected in the first set. This segmentation allows the system to process only necessary candidate words, avoiding the exponential growth of processing all possible candidate words when multiple LRPs are used.
Solution Approach 2:
The patent applies partial action by processing only a subset of candidate words rather than all possible candidate words. When errors are detected using the first set of LRPs, the system generates candidate words only for the second set of LRPs; otherwise, it stops processing. This partial processing approach maintains error correction capability while significantly reducing the total number of candidate words that need to be evaluated.
2Reliability
If the quantity of least reliable positions (LRPs) is increased to correct more errors, then the error correction capability is improved, but the processing time increases due to more candidate words to evaluate
Solution Approach 1:
The patent performs preliminary error detection using the first set of LRPs before generating candidate words for the second set of LRPs. This preliminary action allows the system to determine early whether error correction is needed, avoiding unnecessary processing time when no errors are present. The conditional processing structure ensures that time-consuming candidate word generation occurs only when necessary.
Solution Approach 2:
The system performs partial processing by evaluating only the necessary subset of candidate words based on error detection results. This approach maintains the ability to correct multiple errors while significantly reducing the average processing time compared to evaluating all possible candidate words generated from multiple LRPs.
3Reliability
If the quantity of least reliable positions (LRPs) is increased to improve error identification, then the error identification capability is improved, but the memory usage increases due to storing more candidate words
Solution Approach 1:
The patent segments candidate word generation into two stages: first generating candidate words from the first set of LRPs, then conditionally generating candidate words from the second set of LRPs only when errors are detected. This segmentation reduces the peak memory usage by avoiding the simultaneous storage of all possible candidate words that would result from processing all LRPs unconditionally.
Solution Approach 2:
The system stores only the necessary subset of candidate words required for error correction rather than all possible candidate words. By conditionally generating and storing candidate words based on error detection results, the system maintains adequate memory for error correction while significantly reducing overall memory consumption.
Data Source
AI summary
A system is configured to receive a word on which to perform forward error correction; identify least reliable positions that correspond to encoded bits, within the word, associated with a lowest level of reliability; generate candidate words based on different combinations of inverted encoded bits; identify a pair of candidate words that includes a candidate word and another candidate word, the candidate word includes an inverted most reliable bit of the encoded bits within the candidate word; identify a quantity of errors within the candidate word; determine whether the quantity of errors corresponds to an odd value; invert a parity bit associated with the candidate word when the quantity of errors corresponds to the odd value; select the other candidate word when the parity bit is inverted; and perform forward error correction, on the word, using the other candidate word based on selection of the other candidate word.