HARQ Soft-Bit Storage Using MSB Extraction and Scaling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Hybrid automatic retransmission request (HARQ) techniques require large memory buffers to store soft information for optimal retransmission combining, leading to performance loss or increased memory bandwidth requirements, presenting an undesirable design choice between smaller buffers for performance loss or larger buffers for improved performance.
Innovation Solution
Implementing HARQ combining techniques that store only a subset of soft-bits generated for each transmission, allowing for increased soft-bits per symbol to enhance receiver performance while reducing memory requirements, and applying different scaling factors to improve dynamic range and precision.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the same number of soft-bits that are used for FEC are also stored in memory for HARQ combining purposes, then optimal retransmission combining performance is achieved, but large external memory and/or increased memory bandwidth requirements are needed
Solution Approach 1:
The patent extracts only the most significant bits (MSBs) of the soft information for storage in memory, while keeping the full precision soft information in registers for immediate decoding operations. This selective extraction of critical bits reduces memory bandwidth requirements while preserving essential information for HARQ combining performance.
Solution Approach 2:
The patent applies different precision levels to different storage locations: full precision is maintained in registers for active processing, while reduced precision (MSBs only) is stored in memory for long-term retention. This local differentiation of quality allows optimal performance where needed while reducing overall memory burden.
2Quantity of substance
If a smaller number of soft-bits per symbol are stored, then a smaller buffer is required, but performance loss occurs
Solution Approach 1:
The patent segments the soft information into two parts: the most significant bits (MSBs) that capture the essential reliability information are stored in memory, while the less critical bits are kept in registers or discarded. This segmentation allows the buffer to be smaller while preserving the critical information needed for decoding performance.
Solution Approach 2:
The patent stores only a partial representation (MSBs) of the soft information in memory, which is sufficient for HARQ combining operations. This partial action approach avoids storing redundant information while maintaining adequate performance through the retained significant bits.
Data Source
AI summary
Methods and systems for decoding a data packet using soft-bit retransmission combining are provided herein. A first transmission of the data packet is received and corresponding soft information is generated. The soft information is represented by a first set of bits and a subset of those bits is stored. Subsequently, a second transmission of the data packet is received and corresponding soft information is generated. The soft information corresponding to the second transmission is represented by a second set of bits and combined with the stored subset of soft-bits corresponding to the first transmission to produce a third set of bits. At least a portion of the data packet is decoded based on the third set of bits.