基于改进滑窗迭代DFT的动态脉率变异性提取

Dynamic pulse rate variability extraction method based on improved sliding window iterative DFT

针对现有脉率变异性(PRV)提取方法对噪声、采样频率敏感,计算量大等不足,提出一种可从动态脉搏信号基波分量中提取PRV信号的方法。通过对滑窗迭代DFT进行简化以及自适应调整滑动窗口的宽度,提高计算脉搏信号基波的速率和准确性。同时,采用动态差分阈值与人工检测相结合的方法提取PRV信号,作为标准分析所提出方法的准确性。将所提出方法用于提取不同采样频率、受不同噪声污染以及被试者休息、视觉疲劳、心律不齐等状态下脉搏信号的PRV信号;设计实验,验证其提取动态脉率变异性(DPRV)信号的准确性和实时性。结果表明在不同采样频率、不同噪声水平以及被试者处于不同状态下,该算法仍然保持着很高的准确性;并可以准确地提取动态脉率变异性信号。

Aiming at the defects that the existing pulse rate variability （PRV） extraction methods are sensitive to noise and sampling fre- quency, and require heavy computation, a new method is proposed to extract pulse rate variability （PRV） from the fundamental compo- nent of dynamic PhotoPlethysmoGraphy （ PPG ） signal. Through simplifying the sliding window iterative Discrete Fourier Transform （DFT） and adaptively adjusting the sliding window width, the computation speed and accuracy of fundamental component of the dynamic pulse rate signal are improved. Furthermore, the dynamic difference threshold and manual detection method are combined to extract the PRV signal, which is used as a criterion to analyze the accuracy of the proposed method. The proposed method was used to extract the PRV signals of the PPG signals under the conditions of various sampling frequencies and various noise levels, as well as when the sub- jects are in different states （rest, visual fatigue, arrhythmia and etc. ）. Experiments were designed to verify the accuracy and real-time performance of this method in extracting the Dynamic Pulse Rate Variability （DPRV） signals. The results show that the proposed method still has high accuracy under the conditions of different sampling frequencies, different noise levels as well as when the subjects are in different states; and the proposed method can extract the dynamic pulse rate variability （DPRV） accurately and in real-time.