Dual List Decoding of Convolutional Codes for Short Messages
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing digital communication systems face challenges in efficiently transmitting and decoding short block-length messages, as capacity-approaching codes like LDPC and Turbo codes perform poorly at short block-lengths, and conventional decoders like Viterbi and BCJR algorithms have high latency and complexity.
Innovation Solution
The use of dual list decoding of convolutional codes, where messages are encoded using a combination of two convolutional codes with expurgating linear functions, and decoded using separate partial decoders to reconstruct the message bits, with a fallback to regular Viterbi decoding when necessary, to achieve lower latency and complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If capacity-approaching codes like LDPC and Turbo codes are used, then coding efficiency approaches the Shannon limit, but performance deteriorates at short block-lengths
Solution Approach 1:
The patent segments the received codeword into multiple sub-blocks and processes each sub-block independently through separate decoders. This segmentation allows the system to achieve better performance at short block-lengths by treating each sub-block as a smaller, more manageable unit while maintaining the overall coding efficiency through the structured combination of multiple convolutional codes.
2Measurement precision
If conventional decoders like Viterbi and BCJR algorithms are used, then decoding accuracy is maintained, but latency and complexity increase
Solution Approach 1:
The patent divides the decoding task into multiple parallel sub-tasks, with each sub-block processed by a separate decoder. This segmentation reduces the complexity of each individual decoder compared to processing the entire block through a single Viterbi or BCJR algorithm, while maintaining decoding accuracy through the structured combination of results from multiple decoders.
Solution Approach 2:
The patent employs multiple partial decoders that each process a portion of the received data rather than a single decoder processing the entire block. This partial action approach reduces the computational burden on each decoder, lowering overall complexity and latency while maintaining accuracy through the coordinated combination of partial decoding results.
3Measurement precision
If conventional decoders like Viterbi and BCJR algorithms are used, then decoding accuracy is maintained, but latency increases
Solution Approach 1:
The patent segments the received codeword into multiple sub-blocks that can be processed in parallel by separate decoders. This segmentation enables concurrent processing of multiple sub-blocks, significantly reducing the overall decoding latency compared to sequential processing through a single Viterbi or BCJR algorithm, while maintaining decoding accuracy through the structured combination of results.
Solution Approach 2:
The patent uses multiple partial decoders to process different portions of the received data simultaneously. This partial action approach enables parallel processing that reduces the total decoding time and latency, while the coordinated combination of partial results maintains the required decoding accuracy.
Data Source
AI summary
Communication systems and methods are disclosed involving the transmission and reception of short block-length messages. In many embodiments, transmitters are utilized that encode messages by combing: at least a first partial codeword generated using a first encoder based upon a first convolutional code and a first expurgating linear function; and a second partial codeword generated by a second encoder based upon a second convolutional code and a second expurgating linear function. In a number of embodiments, receivers are utilized that decode received messages using a full decoder that reconstructs the transmitted message bits based upon: at least a first set of partial codewords produced by a first partial decoder based upon the first convolutional code and the first expurgating linear function; and a second set of partial codewords generated by a second partial decoder based upon the second convolutional code and the second expurgating linear function.