Polar Decoder Soft-Information Exchange for HARQ Retransmissions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing approaches to decoding polar codes over multiple transmissions in HARQ systems are suboptimal due to limited block error performance, as they do not account for information from previous transmissions and exchange hard bits without considering reliability, leading to inefficient use of aggregated block lengths.
Innovation Solution
A method where soft information, such as probabilities or log-likelihood ratios, is exchanged among polar decoders across different transmissions to facilitate iterative decoding, allowing each information bit to benefit from all relevant transmissions rather than just the most recent one.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If hard bits are exchanged among polar decoders without considering reliability, then the decoding process is simple, but the block error performance is suboptimal
Solution Approach 1:
The patent changes the parameter of information exchange from hard bits to soft information (log-likelihood ratios), which carries reliability information. This allows the decoding process to consider the reliability of decoded bits, thereby improving block error performance while maintaining manageable complexity through the use of well-defined mathematical transformations.
Solution Approach 2:
The patent implements an iterative feedback mechanism where soft information from previous transmissions is fed back into the current decoding process. Each polar decoder provides soft information about decoded bits to other decoders, which then use this feedback to improve their decoding accuracy, creating a closed-loop system that progressively reduces block error rate.
2Reliability
If information from previous transmissions is not utilized, then the decoding process is straightforward, but the aggregated block lengths are not effectively used
Solution Approach 1:
The patent merges information from multiple transmissions by combining soft information from previous transmissions with current received signals. The log-likelihood ratios from different transmissions are combined mathematically, effectively aggregating the block lengths and improving the reliability of decoding without requiring separate independent decoding processes.
Solution Approach 2:
The patent performs preliminary decoding of previous transmissions to generate soft information before the current transmission is fully processed. This preliminary action allows the system to prepare reliability information in advance, which is then integrated into the current decoding process, improving overall block error performance while organizing the computational workload efficiently.
3Reliability
If soft information is exchanged among polar decoders, then the block error rate improves, but the computational complexity increases
Solution Approach 1:
The patent uses log-likelihood ratios as the parameter for soft information, which allows for efficient mathematical operations. The use of LLRs enables the system to perform reliability-based decoding with computationally tractable operations, avoiding the need for more complex probabilistic calculations while still achieving improved block error rate through soft information exchange.
Data Source
AI summary
A node (110, 115) receives (804) transmissions associated with a given set of information bits, wherein each of the transmissions use a different polar code and share one or more information bits of the given set of information bits. The node determines (808), at each of a plurality of polar decoders (505, 605) of the node, soft information for each information bit included in an associated one of the transmissions, wherein each of the plurality of polar decoders is associated with a different transmission of the transmissions. The node provides (812), from each polar decoder of the plurality to one or more other polar decoders of the plurality, the determined soft information for any information bits shared by their respective associated transmissions, and uses (816) the provided soft information in an iterative decoding process to decode one or more of the received transmissions.