基于低复杂度编译码的数据流错误纠错方法

摘要

针对单粒子翻转可能带来的数据流错误，设计一种改进的数据流错误纠错方法。利用线性分组码的相关理论，分析常用数据流容错方法的容错能力，从线性分组码的编译码原理出发给出一种低复杂度编译码算法，基于该编码的容错方法能够以较少的开销纠正单粒子翻转造成的单比特数据错误。实验结果表明，该方法能够有效纠正单粒子翻转造成的数据错误，与常用的纠检错方法相比，具有较优的纠错性能和较少的容错开销。%This paper designs a kind of encoding and decoding algorithm with a low complexity based on the data correction method to resolve the data stream errors which Single Event Upset(SEU) may bring. It uses the theory of linear block codes to analyze various methods of data fault tolerance, and designs a kind of encoding and decoding algorithms with a low complexity of linear block code from the encoding and decoding principle of linear block codes,the fault-tolerant coding method can effectively correct single-bit data errors caused by SEU, with low fault-tolerant overhead. Fault injection experiments show that this method can effectively correct data errors caused by single event upset,compared with other common error detection or correction methods,error correction performance of this method is superior,while its fault tolerance cost is less.