LDPC Channel Encoding with Matrix Shortening for Flexible Codeword Lengths
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current LDPC code systems, such as DVB-S2, are limited to only two codeword lengths, which restricts their extendibility and flexibility, and require separate parity-check matrices for each length, leading to inefficient memory usage and inability to support various data rates needed in modern communication systems.
Innovation Solution
A method and apparatus for generating LDPC codes with different codeword lengths using shortening or puncturing techniques, allowing for the creation of multiple codeword lengths from a given LDPC code, specifically by grouping columns in a parity-check matrix, determining a shortening range, and applying a predetermined shortening pattern, thereby supporting various data rates without the need for new parity-check matrices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If separate parity-check matrices are used for each codeword length, then the system can support fixed codeword lengths, but memory efficiency decreases and flexibility is lost
Solution Approach 1:
The patent applies universality by designing a single parity-check matrix that can serve multiple codeword lengths through column group shortening. Instead of creating separate matrices for different codeword lengths, the same base matrix is used with selective column removal based on shortening patterns, making the system adaptable to various data rates while maintaining optimal decoding characteristics.
Solution Approach 2:
The patent segments the parity-check matrix into column groups that can be selectively shortened. By dividing the matrix into manageable column groups and applying shortening patterns to specific groups, the system can efficiently generate different codeword lengths from a single base matrix, improving both flexibility and memory efficiency.
2Adaptability or versatility
If multiple parity-check matrices are stored for different codeword lengths, then various data rates can be supported, but memory requirements increase
Solution Approach 1:
The patent reduces memory requirements by making a single parity-check matrix universal across multiple codeword lengths. Through column group shortening techniques, the same base matrix supports various data rates by selectively removing column groups, eliminating the need to store multiple separate matrices in memory.
Solution Approach 2:
Instead of storing multiple original parity-check matrices, the patent creates shortened versions by copying and selectively removing column groups from a single base matrix. This copying approach with selective deletion maintains the ability to support various data rates while significantly reducing the memory footprint.
3Adaptability or versatility
If column-by-column shortening is applied, then precise codeword lengths can be achieved, but processing complexity increases
Solution Approach 1:
The patent simplifies processing by segmenting the shortening operation into column group level rather than individual column level. By working with pre-defined column groups and applying shortening patterns at the group level, the system achieves precise codeword lengths while significantly reducing the computational complexity compared to column-by-column operations.
Solution Approach 2:
The patent applies preliminary action by pre-defining column groups and shortening patterns before the actual encoding process. These pre-established structures allow the system to quickly generate different codeword lengths by simply selecting and removing predefined column groups, avoiding complex real-time processing during encoding.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
A method for channel encoding in a communication system using a DVB-S2 Low-Density Parity-Check (LDPC) code; the method comprising reading parity check matrix information; determining a length K2, where K2 denotes a length of information bits to be obtained by shortening, wherein the information bits include 168 BCH parity bits; based on the length K2, performing shortening on information bits in an order according to a predetermined shortening pattern defined in a given table.