Optimized hard-decision decoding for generalized product codes in high-speed communication systems

Unmet Need

Communication channels often introduce errors due to noise, leading to the need for robust error control mechanisms. While techniques like forward error correction (FEC) add redundant bits to detect and correct transmission errors, existing decoding processes such as bounded-distance decoding (BDD) can misinterpret correct codewords as errors, leading to miscorrections. Methods like extrinsic message-passing help mitigate this issue, but they require increased data flow and storage, which are not ideal for high-speed systems. There is a need for a more efficient and reliable decoding process to minimize miscorrections effectively while maintaining low complexity.

Technology

Duke inventors have developed a novel iterative decoding algorithm for component-based codes, including Generalized Product Codes (GPCs) and staircase codes. This technology is intended for use in high-speed communication systems, particularly in optical transport networks (OTNs) and other cost-sensitive applications. The algorithm improves bounded-distance decoding by detecting and preventing miscorrections and implementing error backtracking during the decoding process. Simulations show that this approach significantly reduces the error floor and delivers nearly ideal performance, preventing miscorrections and providing an additional coding gain of approximately 0.4 dB at a post-FEC bit error rate, outperforming conventional decoding techniques. The algorithm has been demonstrated through simulations and offers significant improvements with minimal additional development required.

Advantages

• Reduces miscorrection impact: Minimizes undetected errors in component codes.
• Superior performance: Improves decoding accuracy with an additional coding gain of 0.4 dB.
• Low complexity: Maintains efficiency for high-speed systems without increasing data flow or storage requirements.