Sparse Graph Code Decoding with Hybrid Erasure Recovery

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current maximum likelihood decoding methods for sparse graph codes, such as Gauss elimination, incur high computational complexity, limiting their applicability and power efficiency in error correction for erasure channels.

Innovation Solution

Implementing a hybrid decoding method that combines trivial decoding and Gauss elimination, where trivial decoding is used to restore lost data with minimal operations and Gauss elimination is applied selectively to reduce overall computational complexity, thereby enhancing error correction capabilities.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If Gauss elimination method is used for maximum likelihood decoding, then decoding accuracy is improved, but computational complexity increases

Engineering Contradiction:
Improvedecoding accuracyVSAvoidcomputational complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the decoding process into two distinct phases: trivial decoding that handles easily recoverable lost packets using simple substitution, and Gauss elimination decoding that handles remaining complex cases. This segmentation allows the system to achieve maximum likelihood decoding accuracy while minimizing computational complexity by applying the computationally intensive Gauss elimination method only when necessary, rather than to all lost packets.

Inventive Principle:
Principle #1Segmentation

2Reliability

If code length is increased to improve error correction performance, then tolerance to error is improved, but decoding complexity increases

Engineering Contradiction:
Improveerror correction capabilityVSAvoiddecoding complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent applies partial action by using trivial decoding for the majority of lost packet recovery cases, which requires minimal computational resources. Only when trivial decoding cannot resolve the lost packets does the system activate Gauss elimination decoding. This partial application of the more complex algorithm maintains error correction capability for long codes while avoiding the full computational burden that would otherwise be required.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP3035540B1Maximum likelihood erasure decoding of sparse graph codes
Publication Date: 2019.10.09 NIPPON TELEGRAPH & TELEPHONE CORP
  • EP3035540B1 patent drawingFigure 1
  • EP3035540B1 patent drawingFigure 2
  • EP3035540B1 patent drawingFigure 3

AI summary

It is an object of the present invention to make it possible to implement maximum likelihood decoding of a sparse graph code at low computational complexity in the maximum likelihood decoding of the sparse graph code. According to the present invention, in the maximum likelihood decoding of the sparse graph code, a lost data decoding process by a trivial decoding method and a lost data decoding process by a Gauss elimination method are performed repeatedly and alternately.