LDPC Code Shortening Patterns for Variable-Length DVB-S2 Decoding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current LDPC code systems, such as DVB-S2, are limited to only two codeword lengths, restricting their extendibility and flexibility, and require separate storage of independent parity-check matrices for each length, which reduces memory efficiency and limits their ability to support various data rates in communication systems.

Innovation Solution

The method involves generating LDPC codes with different codeword lengths using shortening or puncturing techniques, where column groups in the parity-check matrix are ordered and shortened according to a predetermined pattern, allowing for the inclusion of BCH parity bits and selective exclusion of columns, thereby supporting various codeword lengths without the need for new parity-check matrices.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If independent parity-check matrices are stored for each codeword length, then LDPC codes can support fixed codeword lengths, but memory efficiency decreases and system flexibility is limited

Engineering Contradiction:
ImproveLDPC code performanceVSAvoidmemory storage requirement
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

A single parent parity-check matrix is designed to serve multiple codeword lengths through systematic shortening. The matrix structure incorporates column groups that can be selectively removed to generate different codeword lengths (e.g., 16200, 13320, 10800, 7200, 3240 bits), allowing one matrix to perform the function of multiple matrices and significantly reducing memory storage requirements.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The parent parity-check matrix is divided into multiple column groups, where each group contains a specific number of columns (e.g., 360 columns per group). This segmentation enables flexible shortening by removing integer multiples of column groups to achieve different codeword lengths while maintaining the matrix's structural properties and decoding performance.

Inventive Principle:
Principle #1Segmentation

2Device complexity

If only two fixed codeword lengths are supported, then the system structure remains simple, but adaptability to different data rates is restricted

Engineering Contradiction:
Improvesystem structure complexityVSAvoidsupport for various data rates
Core Design Contradiction:
Device complexityVSAdaptability or versatility

Solution Approach 1:

The system transitions from static fixed-length code support to dynamic variable-length code support. By enabling selective shortening through column group removal, the same parent matrix can adaptively generate codewords of different lengths based on channel conditions and data rate requirements, making the system flexible without requiring separate matrices for each length.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The codeword length parameter is made variable through systematic shortening. The parent matrix with fixed dimensions (e.g., 16200 columns) can generate codewords of different lengths by removing different numbers of column groups, allowing the system to adapt to various data rates while maintaining a single underlying matrix structure.

Inventive Principle:
Principle #35Parameter changes

3Manufacturing precision

If column-by-column shortening is performed, then precise codeword length control is achieved, but processing complexity and time increase

Engineering Contradiction:
Improvecodeword length precisionVSAvoidshortening processing time
Core Design Contradiction:
Manufacturing precisionVSLoss of time

Solution Approach 1:

Instead of processing columns individually, the matrix is segmented into column groups where each group contains a fixed number of columns (e.g., 360). Shortening is performed by removing entire column groups or integer multiples of groups, which maintains precise control over the final codeword length while significantly reducing processing complexity and time compared to column-by-column operations.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP2239854B9Shortening and puncturing of low-density parity-check (LDPC) codes for channel decoding
Publication Date: 2015.02.18 SAMSUNG ELECTRONICS CO LTD
  • EP2239854B9 patent drawingFigure 1~2
  • EP2239854B9 patent drawingFigure 3
  • EP2239854B9 patent drawingFigure 4

AI summary

A method for a channel decoding in a communication system using a Low-Density Parity-Check (LDPC) code, the method comprising: demodulating a signal transmitted from a transmitter; determining a position of a shortened bit by estimating information on a shortening pattern of an LDPC code from the demodulated signal from the demodulating step; and decoding data using the determined position of the shortened bit. Hereby a method for channel decoding supporting various code lengths from a given LDPC code in the DVB-S2 system is provided.