生物电放大电路多通道频谱差异校正算法

Calibration algorithm for spectrum difference among multi-channels of bioelectricity amplifier

多通道生物电放大电路通道间固有的频谱特性差异影响后续模式识别的准确率,针对这一问题,提出了一种基于通道相幅频特性的频谱差异校正算法。首先,分析了多通道间频谱差异的影响,在此基础上给出算法实现原理及步骤:先利用各通道输出信号与其采样信号的傅里叶变换关系确定其频域离散校正区间,并估算该区间上各离散频率点的幅度校正因子和相位校正偏置,进而实施校正运算,最后通过傅里叶逆变换得到校正后的信号。仿真结果表明,该算法能够将通道间频谱差异的不良影响降低到50%以下;将该算法用于人体穴位识别的数据预处理上,结果表明,应用该算法后,穴位识别准确率得到显著提高。

The inherent spectrum difference among multi-channels of bioelectricity amplifier impacts the accuracy of subsequent signal recognition.Aiming at this problem,a spectrum difference calibration algorithm is presented based on the amplitude and phase frequency characteristics.First,the impact of the spectrum difference among multi-channels is analyzed,and then the principle and procedure of the algorithm are given.The Fourier transform relationships between the continuous signals and sample sequences of the channels are utilized to determine the discrete calibration interval of the output signals in frequency domain,and estimate the amplitude calibration factor and phase calibration bias at each discrete frequency point in the interval; then the calibration operation is implemented,and the inverse Fourier transform is used to obtain the calibrated signal.Simulation results show that the impact of multi-channel spectrum difference on original signal is reduced to below 50％ using the proposed algorithm.The algorithm was employed in the data preprocessing of human acupoint identification,the results indicate that after using the proposed algorithm the accupoint recognition rate is improved evidently.