Joint Source-Channel LDPC Coding for Flexible 5G NR Rate Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing double LDPC joint source and channel coding scheme for 5G NR systems has limitations in supporting a variety of source coding rates and integrating with existing LDPC-based channel coding, making it difficult to adapt to the 5G NR system effectively.
Innovation Solution
The proposed method determines a source coding rate based on entropy rate and source code length, using a source coding matrix derived from a channel coding basis matrix to perform joint source and channel coding, allowing for a higher number of supported source coding rates and improved flexibility in code rate selection, thereby adapting to the 5G NR system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a fixed source coding matrix is used in double LDPC joint source and channel coding, then the decoding complexity is reduced, but the number of supported source coding rates is limited
Solution Approach 1:
The source coding matrix is segmented into a basis matrix and a lifting matrix. The basis matrix contains the fundamental structure, while the lifting matrix enables rate adaptation through parameter changes. This segmentation allows the system to support multiple source coding rates by modifying the lifting matrix without changing the overall matrix structure, thus maintaining decoding complexity while increasing adaptability.
Solution Approach 2:
The invention uses parameter changes in the lifting matrix to adapt the source coding matrix to different source coding rates. By changing parameters such as the lifting size and circulation shift values in the lifting matrix, the system can support multiple source coding rates (e.g., 1/4, 1/3, 2/5, 1/2, 3/5, 2/3, 3/4) without redesigning the entire coding matrix, thereby resolving the contradiction between adaptability and complexity.
2Reliability
If independent source and channel coding are used, then the system design is simpler, but the bit error rate performance is worse
Solution Approach 1:
The invention merges source coding and channel coding into a joint coding scheme where the source coding matrix is derived from the channel coding basis matrix. This merging allows the system to exploit the correlation between source and channel coding processes, improving bit error rate performance by enabling joint decoding that simultaneously optimizes both source reconstruction and error correction, while maintaining manageable complexity through the structured matrix derivation approach.
3Adaptability or versatility
If the source coding output does not match channel coding input bits, then the coding scheme is more flexible, but it cannot be integrated into existing 5G NR systems
Solution Approach 1:
The invention creates a universal source coding matrix that can serve multiple functions: it is compatible with existing 5G NR channel coding frameworks while simultaneously supporting multiple source coding rates. The basis matrix structure is designed to be universal, allowing it to work with standard 5G NR channel codes, while the lifting matrix provides the flexibility needed for different source coding rates, thus achieving both compatibility and operational flexibility.
Data Source
AI summary
This application relates to an information processing method and device. A first device determines a first source coding rate. The first device determines a source coding matrix based on the first source coding rate and a channel coding basis matrix. The first device inputs a source sequence into the source coding matrix, to perform source coding on the source sequence. In embodiments of this application, the source coding matrix may be determined based on the channel coding basis matrix, which is equivalent to determining source coding information based on channel coding information, to implement joint source and channel coding. Through the joint source and channel coding, a quantity of output bits of source coding can match a quantity of input bits of channel coding, so that technical solutions in embodiments of this application can be adapted to a 5G NR system.