CAJ | 학술논문

虽然传统最小均方（ LMS）算法具有结构简单、计算量小和稳定性好等优点，但其存在收敛速度与稳态误差相矛盾的问题。归一化LMS算法是用输入信号的功率对步长进行归一化，是一种重要的变步长算法，但其使用的全局步长因子是固定的。为了解决这一问题，将该算法与基于正弦函数的变步长LMS算法相结合，提出了一种基于正弦函数的归一化变步长LMS算法。仿真结果表明，该算法优于传统LMS算法、归一化LMS算法和基于正弦函数的变步长LMS算法。
수연전통최소균방（ LMS）산법구유결구간단、계산량소화은정성호등우점，단기존재수렴속도여은태오차상모순적문제。귀일화LMS산법시용수입신호적공솔대보장진행귀일화，시일충중요적변보장산법，단기사용적전국보장인자시고정적。위료해결저일문제，장해산법여기우정현함수적변보장LMS산법상결합，제출료일충기우정현함수적귀일화변보장LMS산법。방진결과표명，해산법우우전통LMS산법、귀일화LMS산법화기우정현함수적변보장LMS산법。
Although the traditional LMS algorithm has simple structure,small burden in computation,good stability and so on, there is contradiction in relationship between steady-state error and convergence.The normalized LMS algorithm is to normalize the step-size by using the power of the input signal,and it is an important variable step-size algorithm,but the used global step size factor is fixed.In order to solve this problem,the algorithm is combined with the variable step-size LMS algorithm based on sine function,and a new normalized variable step-size LMS algorithm based on sine function is proposed.The computer simulation results show that this new algorithm is better than traditional LMS algorithm,normalized LMS algorithm and variable step-size LMS algorithm based on sine func-tion.