Polar Code Bit Selection with Nested Quantized Reliability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing polar code encoding solutions require high storage overheads due to the need for large quantities of mother code sequences to support various code lengths and rates, which complicates the determination of the information bit set and increases system storage requirements.
Innovation Solution
The method involves obtaining a target quantized sequence based on a basic quantized sequence to determine a non-fixed bit position set, facilitating polar code encoding or decoding, thereby reducing storage overheads by simplifying the process of determining the information bit set and reducing the need for extensive storage of mother code sequences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a large quantity of mother code sequences are stored to support all combinations of code lengths and code rates, then the system can support various coding requirements, but the storage overheads of the system become very high
Solution Approach 1:
The patent segments the codebook into multiple codebook groups, where each group corresponds to a specific code rate. Instead of storing all possible code sequences for all code rates and lengths, the system only needs to store a representative set for each code rate group. The actual code sequence can be determined by combining the group identifier with an index within that group, significantly reducing the total storage requirements while maintaining support for various coding requirements.
2Ease of manufacture
If offline computation and storage manner is adopted for determining the information bit set A, then the determination can be pre-calculated, but the encoder and decoder need to prestore a large table of correspondences between mother code sequences and code lengths and code rates
Solution Approach 1:
The patent extracts only the essential information needed for encoding from the full correspondence table. Instead of storing complete mappings between all mother code sequences and their parameters, the system extracts and stores only the codebook group identifiers and intra-group indices. The actual code sequence information can be reconstructed on-demand using the polar code generation formula, eliminating the need to store large correspondence tables while maintaining the benefits of pre-computation.
3Device complexity
If the information bit set A is determined through simple calculation, then the encoding process is simplified, but the information bit set A may include check bits or other bits that facilitate decoding
Solution Approach 1:
The patent implements a dynamic approach where the information bit set A is not fixed but can be flexibly determined based on the selected codebook group and index. The system uses a dynamic mapping mechanism that adapts the composition of information bits and check bits according to the specific coding requirements. This allows the determination process to remain relatively simple while ensuring that the appropriate bits are allocated for both information and decoding facilitation purposes.
Data Source
Figure 1
Figure 1a
Figure 2
AI summary
Embodiments of the present invention disclose a polar code encoding and decoding method in a communications system. The method includes: obtaining a basic quantized sequence, where the basic quantized sequence includes a quantized value used to represent reliability corresponding to a polarized subchannel; obtaining a target quantized sequence based on the basic quantized sequence, where a relative magnitude relationship between elements in the target quantized sequence is nested with a relative magnitude relationship between elements in the basic quantized sequence; determining K largest quantized values in the target quantized sequence based on a non-fixed bit length K, and using polarized subchannels corresponding to the K largest quantized values, as a non-fixed bit position set; and performing polar code encoding or decoding based on the non-fixed bit position set.