LDPC Decoding with Serial Message Passing for Faster Convergence
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional LDPC decoders employing a flooding schedule require a high number of iterations for decoding, leading to increased decoding time and memory usage, and have complex convergence checking mechanisms.
Innovation Solution
A method for decoding low-density parity-check codewords using a serial schedule, where messages are calculated iteratively according to a message passing computation rule, reducing the number of iterations and hardware requirements, and simplifying convergence testing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If a flooding schedule is used for LDPC decoding, then all check nodes are updated simultaneously in parallel, but the number of iterations required for convergence increases significantly
Solution Approach 1:
The patent segments the check nodes into multiple groups and processes them in a serial schedule where one group is updated at a time rather than all simultaneously. This segmentation allows the decoder to make progress toward convergence with fewer iterations while maintaining manageable hardware complexity.
Solution Approach 2:
The patent implements periodic action by cycling through different groups of check nodes in successive iterations. Each group is fully processed before moving to the next group, creating a periodic update pattern that achieves convergence faster than traditional flooding while avoiding the need for all nodes to update simultaneously.
2Extent of automation
If a flooding schedule is used for LDPC decoding, then parallel processing is enabled, but memory size requirements increase
Solution Approach 1:
The patent segments the check nodes into multiple groups that can be processed sequentially. This segmentation allows the use of smaller memory structures since not all check node messages need to be stored simultaneously, reducing the overall memory size requirement while still achieving effective decoding.
Solution Approach 2:
The patent applies partial action by updating only a subset of check nodes in each iteration rather than all check nodes. This partial updating approach reduces the memory burden at any given time while still making progress toward convergence, effectively trading some parallelism for reduced memory requirements.
3Reliability
If a flooding schedule is used for LDPC decoding, then convergence can be achieved, but the convergence checking mechanism becomes complex
Solution Approach 1:
The patent extracts the convergence checking function from a complex centralized mechanism and distributes it across the check node groups. Each group can independently track its own convergence status, simplifying the overall convergence checking mechanism while maintaining reliable detection of decoding convergence.
Solution Approach 2:
The patent implements self-service by enabling each check node group to autonomously determine when it has converged based on local information. This self-service approach eliminates the need for complex centralized convergence checking while ensuring reliable detection of decoding completion.
Data Source
AI summary
A method for decoding a noisy codeword (y) received from a communication channel as the result of a LDPC codeword (b) having a number (N) of codeword bits is disclosed. Each codeword bit consists of k information bits and M parity check bits. The product of the LDPC codeword b and a predetermined (M×N) parity check matrix H is zero (H*bT=0) wherein the parity check matrix H represents a bipartite graph comprising N variable nodes (V) connected to M check nodes (C) via edges according to matrix elements hij of the parity check matrix H.—The method comprises receiving the noisy LDPC codeword (y) via said communication channel and calculating for each codeword bit (V) of said transmitted LDPC codeword (b) an a-priori estimate (Qv) that the codeword bit has a predetermined value. The method also comprises calculating iteratively messages on all edges of said bipartite graph according to a serial schedule and a message passing computation rule.


