Neural Network SIC Decoding Policy for Wireless HARQ
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for improving HARQ performance, such as brute-force decoding and large buffer allocation, lead to memory issues and increased decoding complexity and delay in wireless communication systems.
Innovation Solution
A method using a neural network trained with reinforcement learning to determine a decoding policy for successive interference cancellation (SIC) in wireless communication systems, which includes managing HARQ buffers based on log likelihood ratio (LLR) values and interference relationships.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If brute-force decoding is used to improve HARQ performance, then decoding reliability is improved, but device complexity and processing delay increase
Solution Approach 1:
The patent segments the decoding process into multiple stages based on interference levels. Instead of attempting to decode all codewords simultaneously with equal effort, the system divides codewords into groups based on their interference characteristics and processes them in a structured sequence, reducing overall complexity while maintaining reliability
Solution Approach 2:
The patent performs preliminary actions by pre-calculating and storing soft values (LLR values) of successfully decoded codewords in a buffer before attempting to decode interfering codewords. This preliminary storage of interference information allows subsequent decoding attempts to proceed with reduced interference, improving success rates without requiring exhaustive brute-force search
2Reliability
If large buffer allocation is used to improve HARQ performance, then decoding reliability is improved, but memory requirements increase
Solution Approach 1:
The patent applies local quality by allocating buffer resources selectively based on the specific needs of each codeword decoding scenario. Instead of uniformly allocating large buffer capacity for all possible decoding combinations, the system dynamically determines buffer requirements based on the number of codewords, their modulation and coding schemes, and interference relationships, thereby reducing overall memory requirements while maintaining performance
Solution Approach 2:
The patent employs partial action by storing only the necessary soft values in the buffer - specifically, the soft values of successfully decoded codewords that are needed for subsequent interference cancellation. Rather than storing all possible decoding states or redundant information, the system stores precisely the minimal required information to enable successful SIC-based decoding
3Reliability
If all decoding cases are attempted to improve HARQ performance, then decoding reliability is improved, but processing delay increases
Solution Approach 1:
The patent performs preliminary decoding attempts in a structured sequence based on interference levels. By pre-identifying and decoding low-interference codewords first, and storing their soft values for later use, the system eliminates the need to attempt all possible decoding combinations, thereby reducing processing delay while maintaining high reliability
Solution Approach 2:
The patent introduces dynamic adaptation in the decoding process by adjusting the decoding sequence and buffer management strategies based on real-time channel conditions and interference measurements. The system dynamically determines which codewords to decode first and how to manage buffer resources, optimizing the balance between reliability and processing delay for each specific transmission scenario
Data Source
AI summary
Disclosed is a method by which a terminal decodes a codeword in a wireless communication system. Specifically, the method may comprise: receiving a plurality of codewords; and decoding the plurality codewords on the basis of successive interference cancellation (SIC). In particular, the SIC may be performed on the basis of a decoding policy for decoding the plurality of codewords. In particular, the decoding policy may be determined by a neural network trained on the basis of a state and a reward related to the plurality of codewords.