基于改进变步长最小均方算法的心电图信号去噪  

ECG signal denoising using improved variable step size least mean square algorithm

作　　者：张伟[1] 顾旋 梁富娥 吕珊珊 刘东华 ZHANG Wei;GU Xuan;LIANG Fu'e;LÜShanshan;LIU Donghua(College of Information Engineering,Gansu University of Chinese Medicine,Lanzhou 730100,China)

机构地区：[1]甘肃中医药大学信息工程学院,甘肃兰州730100

出　　处：《中国医学物理学杂志》2023年第9期1135-1142,共8页Chinese Journal of Medical Physics

基　　金：甘肃省教育厅创新基金(2022B-113);甘肃中医药大学研究生创新基金(2022CX81)。

摘　　要：针对定步长和现有变步长最小均方(LMS)自适应滤波算法对心电图(ECG)信号去噪效果较差的问题,提出一种基于分式函数改进的变步长LMS算法。首先利用分式函数构造出改进的变步长LMS算法的步长函数式,通过理论和仿真分析得到该算法参数的最佳取值,并在相同条件下与定步长和其他变步长LMS算法进行性能对比,验证了该算法具有更快的收敛速度、更低的稳态误差和更小的计算复杂度。然后将该算法与定步长和其他变步长LMS算法在相同条件下对含多种真实噪声的ECG信号进行去噪效果对比。实验结果表明,相比于其他算法,该算法能更好地去除ECG信号的噪声,且去噪后ECG信号的信噪比最大,均方误差最小,Pearson相关系数最接近于1。An improved variable step size least mean square(LMS)algorithm based on fractional function is proposed to solve the problem of the poor denoising performances of the fixed step size and the existing variable step size LMS adaptive filtering algorithms on electrocardiogram(ECG)signals.The fractional function is used to construct the step function of the improved variable step size LMS algorithm,and the optimal parameter values of the algorithm are obtained through theoretical and simulation analyses.The performance comparison with the fixed step size and other variable step size LMS algorithms under the same conditions verifies that the proposed algorithm has faster convergence speed,lower steady-state error and less computational complexity.Moreover,the proposed algorithm is compared with the fixed step size and other variable step size LMS algorithms on ECG signals containing multiple real noises under the same conditions.The experimental results show that compared with the other algorithms,the proposed algorithm can better remove the noise in ECG signals,and that the denoised ECG signals have the largest signal-to-noise ratio and the minimum mean square error,with a Pearson correlation coefficient closest to 1.

关 键 词：心电图信号 噪声 最小均方算法 收敛速度 稳态误差 

分 类 号：R318[医药卫生—生物医学工程] TN911.7[医药卫生—基础医学]