一种自适应抑制工频干扰的新方法

A New Technique for Adaptive Rejection of Power-Line Interference

在电子测量系统中,被检测的信号有时叠加有某种周期信号的干扰,如最常见的工频干扰。这类干扰信号的频率若落在有用信号的频谱范围之内,用频域处理方法则很难将其单独消除。在这种情况下,自适应滤波技术是一种有效的方法,它能自动跟踪干扰信号的相位和幅值,使干扰成分得到有效抑制而不损害有用信号。本文在自适应技术基础上,进一步提出一种抑制周期信号干扰的新方法,即以自适应算法综合一组离散傅立叶级数（Discrete Fourier Series,DFS）,用以抵消采样序列中的周期性干扰。这种方法可应用于工业或生物医学测量系统中消除工频干扰和其他周期性或准周期性干扰。

A new technique based on the adaptive principle is proposed to deal with power-line interference.Periodical interference can be represented in a good approximation with a discrete Fourier series (DFS). The coefficients of DFS can be calculated by the least-mean-square (LMS) algorithm and used to synthesize a signal that is capable of eliminating interference. Two practical examples are given.Compared with ordinary adaptive filtering techniques, the method proposed possesses the following merits: (1) Less calculating workload is required and real-time processing is easily implemented; (2) no additional reference signal is required; hence simpler hardwares; (3) to reject interference with higher order harmonics, higher order DFS can be used, and periodical interference with any waveform can be eliminated.