Multi-State LDPC Bit Flipping for Better Error Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing LDPC decoding methods are inefficient due to their reliance on two-state systems, which limit error correction capabilities and require significant processor and memory resources when handling soft information.
Innovation Solution
The implementation of multi-state bit flipping decoders with variable and check nodes, allowing for iterative processing and updating of node values based on received indications, enabling the use of three or four states to improve decoding accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If two-state systems are used for LDPC decoding, then device complexity is reduced, but error correction capability deteriorates
Solution Approach 1:
The patent changes the parameter of bit states from two-state (0, 1) to multi-state (at least three states including 0, 1, and erasure state). This allows the decoder to represent uncertain or unreliable bits with an erasure state, improving error correction capability by providing more nuanced information about bit reliability without significantly increasing device complexity.
2Measurement precision
If soft information is used for decoding, then decoding accuracy is improved, but processor and memory resource requirements increase
Solution Approach 1:
The patent discretizes soft information into multi-state hard decisions (0, 1, erasure) rather than using continuous soft information values. This quantization approach maintains decoding accuracy by preserving reliability information through the erasure state while dramatically reducing processor and memory resource requirements by working with discrete states instead of continuous probabilities.
Solution Approach 2:
The patent uses simple discrete state values (0, 1, erasure) that can be stored and processed with minimal resources, replacing complex soft information representations. These discrete states act as simplified proxies that capture essential reliability information without requiring significant computational or memory resources.
Data Source
AI summary
Systems and methods are provided for decoding data using hard decisions and soft information. In particular, the systems and methods described herein are directed to decoders having variable nodes and check nodes, each with multiple states. The systems and methods include receiving, at a decoder during a first iteration, values for each of a plurality of variable nodes, and determining, during a second iteration, one or more indications for each of a plurality of check nodes based on the one or more values of the variable nodes received during the first iteration. The methods further include updating, at the decoder during the second iteration, the values for each of the variable nodes based on the values of the respective variable node received during the first iteration, and the indications for each of the plurality connected check nodes during the first iteration.