LDPC Check Node Grouping for Faster MAP Decoding Convergence
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional LDPC code decoding methods exhibit slow convergence speed due to the two-phase message passing process, which limits communication performance in latency-critical applications like wireless communication systems.
Innovation Solution
A decoding method that jointly updates check nodes using the Maximum-A-Posteriori (MAP) probability process, grouping check nodes and updating them together without updating internal parity nodes, while still maintaining two phases for trellis calculation and information node updates, to enhance convergence speed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If conventional two-phase message passing process is used for LDPC decoding, then implementation is straightforward, but convergence speed is slow
Solution Approach 1:
The patent combines multiple check node updates into a single joint MAP probability calculation step. Instead of updating check nodes sequentially in two phases, the invention jointly processes groups of check nodes connected to the same variable node, merging their update operations into one unified computation that achieves faster convergence while maintaining manageable complexity.
Solution Approach 2:
The patent segments the set of check nodes into groups based on their connectivity to variable nodes. By dividing check nodes into manageable groups that can be jointly processed, the invention enables parallel computation of MAP probabilities for multiple check nodes simultaneously, thereby improving convergence speed without overwhelming system complexity.
2Productivity
If joint MAP update of check nodes is performed, then convergence speed improves, but computational complexity increases
Solution Approach 1:
The patent segments check nodes into groups based on their connectivity patterns to variable nodes. This segmentation allows the joint MAP update to be performed on manageable subsets of check nodes rather than all check nodes simultaneously, reducing the computational burden while still achieving the convergence speed benefits of joint processing.
Solution Approach 2:
The patent applies joint MAP updating to specific groups of check nodes rather than uniformly to all check nodes in every iteration. By selectively applying the computationally intensive joint update operation only where beneficial (to groups connected to the same variable nodes), the invention achieves improved decoding efficiency while controlling overall computational complexity.
Data Source
AI summary
The method includes defining from all the check nodes at least one group of check nodes mutually connected through at least one second variable node defining an internal second variable node. The method includes performing for each group the joint updating of all the check nodes of the group via a Maximum-A-Posteriori (MAP) type process, and the updating of all the first variable nodes and all the second variable nodes connected to the group except the at least one internal second variable node. The method may include iteratively repeating the updates.