在LDPC码的译码算法中,和积算法性能最优但复杂性较高,最小和算法实现简单但性能与和积算法相差较多。针对这一性能与复杂度的矛盾,带有修正项的最小和算法成为研究的热点问题。文中基于一种性能与和积算法接近的修正最小和算法进行研究,对修正项的修正方式进行了简化,简化后的算法在性能上与和积算法仍非常接近,实现复杂度却比原修正最小和算法有明显的降低。
Among the decoding algorithms for low - density parity - check ( LDPC ) code, sum - product algorithm (SPA) has the optimal performance but high complexity, while min- sum algorithm has simple implementation complexity but worse performance. Considering this performance -complexity contradiction, some modified rain - sum algorithms are proposed. In this paper, researches are done based on one of the modified min - sum algorithms that is close to SPA in performance but the way modifying fraction works is simplified. Simplified algorithm remains similar in performance to sum - product algorithm but has a remarkably reduced implementation complexity.