Adaptive filter for filtering power frequency interference in electromyography signal based on EEMD (Ensemble Empirical Mode Decomposition) algorithm

Abstract

The invention discloses an adaptive filter for filtering power frequency interference in an electromyography signal based on an EEMD (Ensemble Empirical Mode Decomposition) algorithm. The adaptive filter is implemented by the following steps: S1, inputting an electromyography signal x(n) with power frequency noise; S2, performing EEMD algorithm decomposition on the electromyography signal x(n) with the power frequency noise through a first system to decompose an IMF (Intrinsic Mode Function) which satisfies IMF definition; S3, reconstructing the IMF obtained in the step S2 into the reference input D(n) of the adaptive filter through a second system; and S4, constructing the adaptive filter by adopting a BLMS (Block Least Mean Square) algorithm through the reference input D(n) reconstructed in the step S3, and outputting a signal e(n) from which the power frequency interference is removed. The adaptive filter has a good adaptive filtering characteristic specific to power frequency interference of different phases and different frequencies, and does not influence the waveform of an original electromyography detection signal basically. The adaptive filter is constructed with the BLMS algorithm, so that the adaptive filter has the advantages of high stability, high computation efficiency, quick convergence and small mean square error.

Description

technical field [0001] The invention relates to the technical field of bioelectricity control, in particular to an adaptive filter based on an EEMD algorithm for filtering power frequency interference in electromyographic signal detection. Background technique [0002] Surface electromyography (SEMG, Surface-Electrimyography) contains the original information of muscle activity in a certain area and a variety of inevitable noises. The above noise seriously affects the detection of SEMG signals, and power frequency interference, as an important interference source in biological signal acquisition, just falls within the main signal frequency range of SEMG signals, and power frequency noise that changes with time, phase and frequency is more difficult to filter out. It is particularly critical to filter out power frequency interference in SEMG signal detection. [0003] In recent years, a variety of methods for filtering power frequency interference have been applied to SEMG s...

AI Technical Summary

Problems solved by technology

[0008] The above-mentioned filtering of power frequency interference has the following disadvantages: (1) Using analog filters for noise reduction, since the accuracy of electronic components will be affected at different temperatures, and electronic components will age to varying degrees over time, so in To a certain extent, it will affect the performance of filtering; (2), although the digital filter using wavelet transform can avoid the shortcomings of the above (1), it has multi-resolution analysis and can deal with sudden changes in signal noise, but once a wavelet is selected for The whole data processing will produce a lot of pseudo-harmonic components, which will reduce the signal-to-noise ratio of the detected signal; moreover, it is not easy to choose a suitable wavelet function; (3), the adaptive filter can adjust the coefficient of the filter The input signal is weighted to generate an output, which is compared with the expected reference or training signal to form an error signal; and the aforementioned adaptive filter tends to converge slowly, with high steady-state error and instability defects

Images