Non-Binary LDPC Decoding with Compressed CW-TMM Sorting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Non-binary LDPC codes in communication systems face high decoding complexity and increased memory usage, particularly in systems like 5G and 6G, which affect hardware efficiency and throughput.
Innovation Solution
A column-wise trellis min-max (CW-TMM) algorithm is employed, incorporating message compression techniques and a 6-stage pipelined structure to reduce sorter complexity and memory usage, optimizing the decoding process for non-binary LDPC codes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If non-binary LDPC codes are used to improve channel capacity and error correction performance, then decoding complexity increases
Solution Approach 1:
The check node processing is divided into column-wise sorting and row-wise sorting stages. The column-wise sorting processes messages from different variable nodes separately, identifying minimum values and their indices. The row-wise sorting then processes these results to generate final check-to-variable messages. This segmentation reduces the complexity of handling non-binary messages by breaking down the processing into manageable sequential steps.
Solution Approach 2:
The patent transforms the message representation from full non-binary values to compressed forms using minimum value extraction and index tracking. Instead of processing complete non-binary messages throughout the decoding process, the system changes parameters by storing only the minimum value and its index, significantly reducing memory requirements and computational complexity while maintaining decoding performance.
2Reliability
If non-binary LDPC codes are used to improve channel capacity, then memory usage increases
Solution Approach 1:
The patent extracts only the essential information from non-binary messages for storage and processing. Instead of storing complete non-binary message vectors in memory, the system extracts and stores only the minimum value and its corresponding index. This extraction principle dramatically reduces memory usage while preserving the necessary information for accurate decoding operations.
Solution Approach 2:
The message compression technique changes the stored parameters from full non-binary message vectors to compact representations consisting of minimum values and indices. This parameter transformation reduces the memory footprint from O(qn) to O(n) where q is the field size and n is the number of variable nodes, making non-binary LDPC decoding feasible in practical hardware implementations.
3Reliability
If traditional decoding algorithms are used for non-binary LDPC codes, then throughput decreases due to high complexity
Solution Approach 1:
The decoding algorithm is segmented into distinct processing stages: column-wise sorting, normalization, row-wise sorting, and message generation. Each stage processes data independently and can be efficiently implemented in parallel for different check nodes or variable nodes. This segmentation enables pipelined hardware architectures that maintain high throughput while handling non-binary messages with reduced complexity operations.
Data Source
AI summary
A method of a first communication node may comprise: storing a compressed channel message in an APP memory, based on a signal received from a second communication node; generating a first V2C message using a first APP message and a first C2V message respectively obtained from the APP memory and a C2V memory, in a VNP process; obtaining a first minimum vector and a second minimum vector by performing CW sorting on the first V2C message, in a CNP process; performing normalization on the first V2C message based on the first minimum vector, in the CNP process; and obtaining a third minimum vector by performing RW sorting on the second minimum vector based on the first minimum vector, in the CNP process.