对于随机信号频率的漂移,如肌电图或超声多普勒,可以通过监视一个或多个功率谱参数的方法进行跟踪。当采用有限长信号来确定这些频率参数时,就会引入随机误差,有时还有系统误差或偏差。就功率谱诸因素而言,已经得到了中频、零交叉频率、转折频率(中值频率是一个特例)的偏差和标准偏差的估计值的近似表达式。以三个人的肌肉在恒力作用下的等张收缩的实验结果为例,从所记录的体表肌电图来看,与本文的理论预估是相符的。在这个例子中,中频的随机误差为最小。这证明,当随机信号幅度分布不是绝对高 Drift of random signal frequencies, such as EMG or Doppler, can be tracked by monitoring one or more power spectral parameters. When using finite length signals to determine these frequency parameters, random and sometimes systematic errors or deviations are introduced. In terms of power spectrum factors, approximations of the estimates of the deviations and standard deviations of the intermediate frequency, the zero-crossing frequency, the corner frequency (the median frequency is a special case), have been obtained. Taking the experimental results of isokinetic contraction of three human muscles under constant force as an example, from the recorded EMG, it is consistent with the theoretical prediction in this paper. In this example, the intermediate frequency has the smallest random error. This proves that when the random signal amplitude distribution is not absolutely high