CAJ | 학술논문

针对 Max-Log-MAP 算法存在译码性能比 Log-MAP 算法差0.5dB,而 Log-MAP 算法复杂度大的问题.论文提出了用线性分段(piecewise-linear term:PLT)的最大值运算来替代Log-MAP算法中雅克比公式,并通过研究比例因子对译码性能的影响,获得了最佳的优化 Turbo 译码性能的方法.研究结果证明,用分段项数 r 等于3,比例因子s为0.7时,获得的Turbo译码性能略优于Log-MAP算法,而算法的计算复杂度比Max-Log-MAP算法仅增加一次乘法、一次加法和求最大值运算.
침대 Max-Log-MAP 산법존재역마성능비 Log-MAP 산법차0.5dB,이 Log-MAP 산법복잡도대적문제.논문제출료용선성분단(piecewise-linear term:PLT)적최대치운산래체대Log-MAP산법중아극비공식,병통과연구비례인자대역마성능적영향,획득료최가적우화 Turbo 역마성능적방법.연구결과증명,용분단항수 r 등우3,비례인자s위0.7시,획득적Turbo역마성능략우우Log-MAP산법,이산법적계산복잡도비Max-Log-MAP산법부증가일차승법、일차가법화구최대치운산.
Aiming at the existing problem that Max-Log-MAP algorithm of decoding performance is worse 0.5dB than Log-MAP algorithm and the Log-MAP algorithm has a greater complexity, the paper presents two methods involving in using the maximizing operation of piecewise-linear term (piecewise-linear term: PLT) to substitute for Jacobian formula in Log-MAP algorithm and introducing the scale factor in extrinsic information so as to optimize Turbo decoding performance. The results show that when the sub-item number r of piecewise-linear term equals 3 and the scale factor s is 0.7, the kind of decoding strategy has slightly better performance than Log-MAP algorithm, as well as the complexity increases only one multiplication, one addition and maximizing operation as compared with that of Max-Log-MAP algorithm.