平面内多模态低频生物力学信号建模与仿真

Modeling and Simulation of in-Plane Multimodal Low-Frequency Biomechanical Signal

由于生物体内肌电信号频率波动较大,现有方法无法局部化分析信号频率,导致低频生物力学信号与实际输入拟合度较差。提出平面内多模态低频生物力学信号建模与仿真。确定最小分辨率,选取最佳光电编码器采集有关肌肉的肌电信号。为消除高频信号干扰,引入频谱分析和幅值转换对数据做预处理;结合局部匹配度与局部方差确定小波空间内融合权重,提取肌肉图像有效信息,实现低频处频率细分目的。将肌电信号通过一体化处理后,计算肌肉活化程度与肌肉力,确定动力学参数,建立动力学信号模型。仿真结果表明,与传统方法相比所提动力学信号模型获取的反馈数据与输入数据更为接近,对生物体共振损伤与防治有着重要临床价值。

In order to avoid the poor fitting between the low frequency biomechanical signal and the actual input caused by the large frequency fluctuation of the EMG signal in vivo, this paper proposes the modeling and simulation of multi-modal low-frequency biomechanical signal in plane. The minimum resolution was determined and the best photoelectric encoder was selected to collect the EMG signals of muscles. The spectrum analysis and amplitude conversion were introduced to preprocess the data for eliminating the interference of high frequency signal. The fusion weight in wavelet space determined by local matching degree and local variance was combined to extract effective information of muscle image, thus completing frequency subdivision at low frequency. After integrating EMG signal, the degree of muscle activation and muscle force were calculated to determine the dynamic parameters and establish the dynamic signal model. The simulation results show that the difference between the feedback data from the dynamic signal model designed in this paper and the input data is smaller, which is of great significance to the prevention and treatment of biological resonance damage.