LDPC Min-Sum Decoding Using Sign Product and Minimum Index
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Min-sum decoding for LDPC codes requires complex circuit configurations due to the need to derive and manage multiple value ratios, particularly the second minimum value, which increases calculation and circuit complexity.
Innovation Solution
A decoding apparatus and method that simplifies min-sum decoding by calculating the total product of signs of prior value ratios before check node processing and using this product to derive extrinsic value ratios, reducing the need for complex calculations and circuitry by identifying and utilizing the minimum value of absolute prior value ratios, and updating extrinsic values accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If min-sum decoding is implemented using conventional methods to derive the second minimum value, then decoding accuracy is improved, but circuit complexity increases due to requiring two memory units and memory swapping operations
Solution Approach 1:
The patent pre-calculates and stores the column index corresponding to the minimum value before check node processing begins. This preliminary action allows the decoding unit to quickly access and compare values during processing without performing complex real-time searches, thereby maintaining decoding accuracy while reducing circuit complexity and avoiding the need for two memory units and swapping operations
Solution Approach 2:
The patent divides the check node processing into two independent parts: (1) retrieving the pre-stored minimum value and its column index, and (2) comparing other values against this minimum. This segmentation allows each part to be implemented with simple logic circuits rather than complex unified processing, reducing overall circuit complexity while maintaining the ability to derive the second minimum value accurately
2Reliability
If conventional min-sum decoding is used requiring derivation of minimum and second minimum values, then decoding performance is improved, but the amount of calculation increases
Solution Approach 1:
The patent performs preliminary identification and storage of the minimum value's column index before the main decoding process. During check node processing, the decoding unit simply compares incoming values against this pre-identified minimum, avoiding repeated searches and complex calculations. This reduces the computational burden during iterative decoding while maintaining the ability to accurately determine both minimum and second minimum values for reliable decoding performance
Solution Approach 2:
The patent creates a simplified representation of the minimum value information (storing only the column index and value) that can be quickly referenced during processing. Instead of maintaining and processing complete value sets, the system uses these copied minimal representations to guide the comparison operations, significantly reducing the amount of calculation required while preserving decoding accuracy
Data Source
AI summary
A min-sum processing unit executes a min-sum algorithm on input data so as to alternately execute check node processing in which an extrinsic value ratio is updated based on prior value ratios and variable node processing in which a prior value ratio is updated based on the extrinsic value ratios. Here, an initializing unit calculates the total product of the signs of the prior value ratios associated with the row to be processed. A deriving unit derives the sign for an extrinsic value ratio associated with the row to be processed, based on the sign of the prior value ratio that is unused in the updating of the extrinsic value ratio and the total product of the signs thus calculated. An updating unit updates an extrinsic value ratio associated with the row to be processed, using the sign thus derived.