Polar Code Rate Matching Using Reliability-Guided Matrix Deletion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing wireless communication technologies, such as Turbo codes and LDPC codes, face challenges in achieving ideal performance for IoT systems due to high complexity and difficulty in achieving flexible code lengths, while polar codes struggle with high implementation complexity for rate matching.
Innovation Solution
A polar encoding and rate matching method that determines to-be-deleted rows and columns in a generator matrix based on a preset mother code length, reducing calculation complexity by using a reliability ranking table to obtain a final generator matrix for encoding and rate matching, eliminating the need for repeated reliability calculations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If polar code rate matching is implemented using traditional methods (exhaust algorithm or greedy algorithm), then the code can achieve flexible code lengths, but the calculation complexity becomes excessively high
Solution Approach 1:
The patent pre-calculates and stores reliability sequences for different code lengths before actual encoding. By performing the complex reliability calculations in advance and storing the results in tables, the system avoids repeating these calculations during real-time encoding operations, thus achieving flexible code lengths while keeping real-time computational complexity low
Solution Approach 2:
The patent prepares reliability sequence tables in advance that cushion against the need for complex real-time calculations. These pre-computed tables serve as a buffer that allows the system to quickly determine optimal puncturing patterns without performing exhaustive searches or complex greedy algorithms during actual encoding
2Adaptability or versatility
If multiple reliability sequence tables are stored for different code lengths, then rate matching for various code lengths can be achieved, but the number of stored tables and memory requirements increase
Solution Approach 1:
The patent designs reliability sequence tables with a universal structure that can serve multiple code lengths. By using a standardized table format and selection mechanism, a single set of tables can support rate matching for various code lengths through appropriate selection and truncation, rather than requiring separate dedicated tables for each code length
Solution Approach 2:
The patent changes the approach from storing multiple complete tables to storing a unified table structure where parameters like code length are varied through selection rather than duplication. The same base table can be adapted to different code lengths by selecting appropriate portions or applying different parameters during the rate matching process
3Reliability
If Turbo code or LDPC code is used to approach Shannon limit with increased code length, then coding performance improves, but encoding and decoding complexity increases significantly
Solution Approach 1:
The patent adopts polar codes which provide a more efficient and simpler encoding/decoding structure compared to Turbo or LDPC codes. The polar code's inherent structure allows it to achieve near-Shannon limit performance with lower computational complexity, effectively replacing the more complex traditional codes with a more efficient alternative
Data Source
AI summary
A polar encoding and rate matching method is provided, including: determining one or more to-be-deleted rows of a generator matrix, where the generator matrix is determined based on a preset mother code length, and a column in which 1 in the one or more to-be-deleted rows is located includes at least one column with a column weight being 1; deleting, from the one or more to-be-deleted rows, a row corresponding to a polarized channel with lowest reliability, and deleting, from a column in which 1 in the row is located, a column with a column weight being 1, to obtain an after-deletion generator matrix; using the after-deletion generator matrix as a generator matrix to repeat the foregoing steps, until an order of a final generator matrix is equal to a target code length; and performing encoding and rate matching on a to-be-encoded bit based on the final generator matrix.