一种基于LS准则和L阶输入矢量的自适应算法

An Adaptive Filtering Algorithm Based On LS Criteria and L-Rank Input Vector

提出了一种新的基于LS准则和L阶输入矢量的自适应滤波算法.该算法采用新的梯度计算公式,使自适应滤波中权矢量的更新比LMS精确和平滑.仿真表明经过250次迭代,新算法就可以收敛于最优值1.6.新算法与基本的解相关LMS算法(DLMS)和LMS算法相比,收敛速度快、稳定性更好.并给出了新算法、DLMS算法和LMS算法性能曲线仿真结果.

A new adaptive filtering algorithm based on LS criteria and L-rank input vector is proposed. The new algorithm uses a new formula to calculate the gradient,and gets a more accurate renewing vector. Simulation shows that the new algorithm converges to the optimal value 1.6,through about 250 calculations. It is much faster and more stable in convergent performance,compared with LMS and DLMS. Simulation in MATLAB is given at last.