一种改进的变步长最小均方算法滤除心电信号运动伪迹的研究

设计一种便携式心电监测装置,其具有低功耗、应用场景宽等优点,能够在运动状态下实时监测人体心电信号(electrocardiogram,ECG)。为了滤除心电信号中的噪声干扰,尤其是运动伪迹(motion artifact,MA)的噪声干扰,在最小均方(least mean squares,LMS)算法的基础上改进步长因子,加快自适应算法的收敛速度,从而保证在最佳权系数附近的失调量最小,且减少权值系数更新的运算量。实验结果表明,算法在处理信号过程中能够得到清晰不失真的原始心电信号,具有运算量小且滤波效果较好等优点。

An improved variable step size LMS algorithm for filtering motion artifacts of ECG signals

This paper designs a portable ECG monitoring device, which has the advantages of low power consumption and wide application scenarios, and can monitor the human body electrocardiogram (ECG) in real time during exercise. In order to eliminate the noise interference in the ECG signal, especially the motion artifact (MA) interference, the optimization step factor is improved based on the traditional LMS algorithm, and the convergence speed of the adaptive process is accelerated, thus ensuring optimization. weight. The amount of misalignment near the coefficient is the smallest, and the amount of calculation of the weight coefficient update is reduced. The experimental results show that the algorithm can obtain clear and undistorted original ECG signals in the real-time processing of signals, which has the advantages of small calculation and good filtering effect.