Hybrid Polar Code Bit Selection for Better Decoding Distance

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Polar codes, particularly in short or medium length, do not exceed the performance of Turbo codes or LDPC codes in decoding, and suffer from error propagation due to hard decision-based bit-by-bit decoding, necessitating improved decoding performance and code distance.

Innovation Solution

A hybrid Polar code is generated by selecting information bits based on both reliability and row weight, optimizing the encoding process to enhance decoding performance and code distance, which is further improved by applying SC or SC-List decoding.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If SC decoding is used for Polar codes, then decoding complexity is reduced to O(Nlog2N), but decoding performance does not exceed Turbo codes or LDPC codes for short to medium code lengths

Engineering Contradiction:
Improvedecoding complexityVSAvoiddecoding performance
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The patent changes the parameter of information bit selection by considering both reliability metrics and row weights from the generator matrix, rather than using traditional reliability-only selection. This dual-criteria approach modifies the code structure parameters to achieve better performance while maintaining SC decoding complexity

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent creates a hybrid Polar code that combines two different selection criteria (reliability and row weight) into a unified code structure, similar to how composite materials combine different properties. This hybrid approach achieves performance closer to Turbo/LDPC codes while maintaining the simplicity of Polar code decoding

Inventive Principle:
Principle #40Composite materials

2Ease of operation

If hard decision is performed after each bit decoding in SC decoding, then decoding process is simplified, but error propagation occurs causing performance degradation

Engineering Contradiction:
Improvedecoding process simplicityVSAvoiddecoding performance
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The patent performs preliminary action by pre-calculating and storing row weights from the generator matrix before the actual decoding process. This pre-computation allows the decoder to use both reliability and structural information without adding complexity during the real-time decoding operation, thereby maintaining simplicity while improving performance

Inventive Principle:
Principle #10Preliminary action

3Reliability

If CRC is cascaded to increase minimum code distance, then error correction capability is improved, but application complexity increases

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

Solution Approach 1:

The patent merges the code distance improvement function into the Polar code structure itself by using hybrid information bit selection based on both reliability and row weights. This integration achieves enhanced error correction capability without requiring an external CRC mechanism, thereby avoiding the additional complexity that would be introduced by cascading CRC

Inventive Principle:
Principle #5Merging (Combining)

Data Source

PatentEP3447925B1Method and apparatus for generating hybrid polar code
Publication Date: 2021.01.27 HUAWEI TECH CO LTD
  • EP3447925B1 patent drawingFigure 1
  • EP3447925B1 patent drawingFigure 2
  • EP3447925B1 patent drawingFigure 3

AI summary

Embodiments of the present invention provide a method and an apparatus for generating a hybrid Polar code. The method includes: obtaining a first matrix of N×N and a sequence that includes N bits, where N is a code length of a hybrid Polar code, N rows of the first matrix correspond to the N bits in the sequence in a one-to-one manner, and N is a positive integer; determining reliability of the N bits, and determining the weight of each row in the N rows of the first matrix; selecting, according to the reliability of the N bits and the weight of the N rows of the first matrix, K bits among the N bits as information bits, or selecting, according to the reliability of the N bits and the weight of the N rows of the first matrix, K rows of the first matrix to construct a second matrix of K×N used for encoding, so as to encode an information bit sequence according to positions of the information bits or according to the second matrix to generate the hybrid Polar code, where K is the length of a to-be-encoded information bit sequence and is not greater than N. In the embodiments of the present invention, not only reliability of a bit is considered, but also the weight of a row that is of a first matrix and corresponds to the bit is considered, so that performance of a Polar code can be improved.