LDPC Decoder Early Termination Using On-the-Fly CRC Checks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Low-density parity-check (LDPC) decoding algorithms face challenges in reducing average iteration number and decoding latency, which affects the throughput of LDPC decoders in error correction systems.
Innovation Solution
The proposed solution involves calculating cyclic redundancy check (CRC) parity bits on-the-fly during LDPC decoding, updating them after each iteration, and terminating the decoding process based on comparison with included CRC bits and LDPC checksum thresholds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If LDPC decoding is performed with multiple iterations to ensure error correction accuracy, then reliability is improved, but decoding latency and average iteration number increase
Solution Approach 1:
The patent performs preliminary CRC parity calculation during the LDPC decoding process itself, rather than after completion. By calculating CRC parity bits on-the-fly during each iteration and comparing with received CRC bits, the system can detect errors early and terminate decoding prematurely when errors are detected, avoiding unnecessary full iterations and reducing overall decoding latency while maintaining reliability
Solution Approach 2:
The patent implements feedback by continuously comparing calculated CRC parity bits with received CRC bits during the decoding process. This feedback mechanism allows the system to monitor decoding progress in real-time, detect errors early, and make dynamic decisions about termination, thereby reducing average iteration number and decoding latency without sacrificing error correction accuracy
2Reliability
If full LDPC decoding iterations are completed to ensure accurate error correction, then reliability is improved, but throughput decreases
Solution Approach 1:
The patent performs preliminary CRC verification during the decoding process by calculating CRC parity bits on-the-fly and comparing them with received CRC bits. This allows early detection of erroneous codewords, enabling the system to terminate decoding for those cases before completing full iterations, thereby increasing average throughput while maintaining reliability for valid codewords
Solution Approach 2:
The patent applies partial action by performing only the necessary portion of LDPC decoding iterations for each codeword. By using CRC-based early termination, the system performs fewer iterations on average for erroneous codewords (excessive action avoided) while ensuring complete decoding for valid codewords (partial action sufficient), optimizing overall throughput without compromising reliability
Data Source
AI summary
Decoding method includes calculating cyclic redundancy check (CRC) parity bits for data on-the-fly; performing a low-density parity-check (LDPC) decoding for the data; if it is determined that an iteration of is finished, updating the calculated CRC parity bits to generate CRC parity bits; comparing the generated CRC parity bits with CRC bits included in the data; and terminating the LDPC decoding based on the comparison result.


