Iterative Decoding Defect Recovery with Branch Metric Suppression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for iterative decoding in data channels fail to reliably recover from defects, as corrupted FIR signals lead to unreliable hard decisions and error propagation, especially in cases of channel defects greater than noise.
Innovation Solution
Stopping iterations at the detector and using the output from the previous stage's outer decoder, or zeroing branch metrics for defective data, to rely on more reliable a priori LLRs from the outer decoder for decoding, thereby discounting defective data and basing results on prior stage outputs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If iterative decoding is performed with multiple stages including soft detector and outer decoder, then decoding accuracy is improved, but error propagation from defective bits worsens reliability
Solution Approach 1:
The patent extracts and removes the harmful component (defective bit contributions) from the iterative decoding process by zeroing out branch metrics for defective bits. This prevents corrupted FIR signal data from propagating errors through the soft detector and outer decoder stages, while preserving the beneficial iterative decoding process for reliable bits.
Solution Approach 2:
The patent applies different processing quality to different parts of the data stream: defective bits have their branch metrics zeroed out (local suppression), while non-defective bits continue to undergo full iterative decoding processing. This localized treatment maintains high decoding accuracy for good bits while preventing error propagation from bad bits.
2Reliability
If channel defects are present causing corrupted FIR signal, then noise handling methods are insufficient, but traditional erasure methods lose information from previous stages
Solution Approach 1:
The patent performs preliminary identification of defective bits before they can corrupt the iterative decoding process. By detecting defects early and zeroing out their branch metrics at the start of processing, the system prevents error propagation while preserving extrinsic LLR information from previous outer decoder iterations for use in subsequent decoding stages.
Solution Approach 2:
The patent utilizes feedback from the outer decoder's extrinsic LLRs back to the soft detector in iterative decoding. By preventing defective bit contributions from corrupting this feedback loop through metric zeroing, the system maintains reliable information flow across iterations while preserving the beneficial feedback mechanism.
3Productivity
If branch metrics from corrupted FIR signal are used in soft detector, then processing continues, but error propagation to adjacent bits occurs
Solution Approach 1:
The patent extracts and removes the harmful component (defective bit contributions) from the iterative decoding process by zeroing out branch metrics for defective bits. This prevents corrupted FIR signal data from propagating errors through the soft detector and outer decoder stages, while preserving the beneficial iterative decoding process for reliable bits.
Solution Approach 2:
The patent applies different processing quality to different parts of the data stream: defective bits have their branch metrics zeroed out (local suppression), while non-defective bits continue to undergo full iterative decoding processing. This localized treatment maintains high decoding accuracy for good bits while preventing error propagation from bad bits.
Data Source
AI summary
In iterative decoding, a data recovery scheme corrects for corrupted or defective data by incorporating results from a previous decoding iteration. In one embodiment, a final multiplexer selects between the final detector output or a previous detector output based on the absence or presence of defective data. In another embodiment, the branch metrics for the defective data, which otherwise would be combined with a priori LLRs from an outer decoder of a prior stage, are ignored so that the a priori LLRs themselves are used alone. The two embodiments can be used together.