CRC-LLR Early Termination in Iterative Detection and Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In digital communication systems like LTE and 5G NR, iterative detection and decoding (IDD) processes consume significant power and resources, with excessive iterations potentially improving symbol quality but also increasing battery drain and memory usage, necessitating a method to determine early termination of IDD to reduce power consumption without compromising quality.
Innovation Solution
The method involves computing log-likelihood ratios (LLRs) for cyclic-redundancy checks (CRC) during IDD, determining if further iterations are likely to fail based on metrics such as mutual information, average LLR magnitude, sign change ratio, or zero LLR count, and terminating the IDD procedure accordingly to conserve resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If iterative detection and decoding (IDD) procedures continue for multiple iterations to improve symbol quality, then the quality of decoded symbols is improved, but power consumption and resource usage increase
Solution Approach 1:
The patent applies preliminary action by computing LLRs for CRC checks before completing full iterative detection and decoding. By evaluating CRC predictions on LLRs early in the process (after initial iterations), the system can determine whether further iterations are likely to succeed before committing additional power resources, thus resolving the contradiction between quality improvement and power consumption
Solution Approach 2:
The patent implements feedback mechanisms by using CRC check results and LLR quality metrics to dynamically control the continuation or termination of IDD procedures. The feedback loop monitors decoding progress and adjusts iteration count based on predicted success probability, allowing the system to maintain symbol quality while avoiding unnecessary power consumption from futile iterations
2Use of energy by moving object
If iterative detection and decoding (IDD) procedures are terminated early to reduce power consumption, then power consumption and memory usage are reduced, but symbol quality may deteriorate
Solution Approach 1:
The system performs preliminary CRC predictions using computed LLRs before terminating IDD procedures. This preliminary assessment ensures that early termination only occurs when success is unlikely, preventing quality deterioration while still achieving power savings by avoiding futile iterations
Solution Approach 2:
The patent dynamically changes the termination parameter (iteration count) based on LLR quality metrics and CRC prediction results. By adjusting the number of iterations according to real-time decoding progress and predicted success probability, the system maintains symbol quality when needed while reducing power consumption when further iterations are unlikely to succeed
3Use of energy by moving object
If log-likelihood ratios (LLRs) are computed for CRC checks during IDD to predict failure, then power consumption is reduced through early termination, but computational overhead increases
Solution Approach 1:
The patent extracts only the essential LLR information needed for CRC predictions rather than processing complete decoded data. By computing LLRs specifically for CRC check purposes and using only those extracted values for termination decisions, the system reduces overall computational overhead while still achieving power savings through early termination
Data Source
AI summary
A system and a method are disclosed for determining early termination during an iterative detection and decoding (IDD) procedure. The method may include computing, during the IDD procedure, one or more log-likelihood ratios (LLRs) of one or more cyclic-redundancy checks (CRC), and determining that at least one of the LLRs predict a failure of a CRC check and, in response, terminating the IDD procedure.