LDPC Matrix Index Storage for Low-Memory Encoding and Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Storing large LDPC parity check matrices efficiently while minimizing storage requirements without compromising encoding/decoding performance is a challenge in communication systems like WiMAX and 802.11n, as existing methods require significant storage volume for effective operation.
Innovation Solution
A data processing system and method that stores indices representing shifting numbers for cyclic permutation matrices, allowing retrieval and use of these matrices to perform LDPC encoding/decoding, thereby reducing storage needs without increasing complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If large LDPC parity check matrices are stored directly, then encoding/decoding performance is maintained, but storage requirements increase significantly
Solution Approach 1:
The patent extracts only the essential information from the full LDPC parity check matrix by storing indices that represent shifting numbers for cyclic permutation matrices. Instead of storing the complete matrix with all its elements, only the critical indices are retained, which can then be used to reconstruct or reference the actual matrix elements during encoding/decoding operations, thereby reducing storage volume while preserving performance
Solution Approach 2:
The patent creates a compressed representation (copy) of the LDPC matrix where indices serve as references to cyclic permutation matrices. Rather than storing duplicate full matrix data, the system stores compact index values that point to or define the actual matrix structures, enabling efficient storage and retrieval without losing the functional characteristics needed for encoding/decoding
2Quantity of substance
If matrix storage is reduced by storing only indices, then storage requirements are minimized, but system complexity increases
Solution Approach 1:
The patent implements a self-service mechanism where the stored indices automatically generate or reference the necessary matrix structures during operation. The system uses the indices to reconstruct cyclic permutation matrices on-demand or to access pre-generated matrices, eliminating the need for manual matrix storage and reduction algorithm complexity. The indices themselves serve the dual purpose of storage and structural definition
3Reliability
If sub-matrix size L is increased to improve code performance, then error correction capability improves, but storage requirements for the parity check matrix increase
Solution Approach 1:
The patent segments the LDPC matrix into cyclic permutation matrices of size L×L, where each matrix is defined by a shifting number index. By storing only the indices that define these segmented matrices rather than the full expanded matrix, the system can use larger sub-matrix sizes for improved error correction capability while keeping storage requirements manageable through the compact index representation
Data Source
AI summary
A data processing system for processing digital data with a low density parity check (LDPC) matrix includes: a storage device for storing a plurality of indices representing a plurality of shifting numbers, where the LDPC matrix comprises an array of elements, and at least one element of the LDPC matrix represents a cyclic permutation matrix that is produced by cyclically shifting columns of an identity matrix to the right according to one of the shifting numbers; and a processing circuit, coupled to the storage device, for retrieving at least one index to recover at least one element of the LDPC matrix according to the index and performing data processing according to the LDPC matrix.