基于改进卡尔曼滤波的列车舒适度平稳性测试仪

Train Comfort and Stability Measuring Based on Adaptive Kalman Filter

为了提高列车舒适度平稳性测量的精确性,针对加速度检测过程中有较多野值干扰的情况,基于活化函数的抗野值算法,结合sage-husa自适应算法设计了改进卡尔曼滤波方法。基于MATLAB设计了仿真实验,实验以美国六级轨道谱模拟值作为列车二自由度模型的输入并加入噪声,得到带野值的列车垂向加速度的模拟值,分别利用改进卡尔曼滤波器、巴特沃兹低通滤波器、普通卡尔曼滤波器对加速度值进行滤波。结果表明,改进卡尔曼滤波的最大均方根误差较巴特沃兹低通滤波与普通卡尔曼滤波相比分别减少了96%和93%。通过在高铁线路上测试实验证明,测试仪可对加速度值进行精确测量并对舒适度平稳性进行精确计算,同时具有更好的便携性。

In order to improve the accuracy of train comfort and stability measurement,in view of the many outliers interference during acceleration detection,an improved kalman filtering method is designed,which combines the sage-husa adaptive algorithm and the outliers rejecting algorithm based on activation function.A simulation experiment is then designed based on the MATLAB.The simulation value of the vertical acceleration of the train with the wild value is obtained by using the analog value of the six-stage track spectrum of the United States as the input of the two-degree-of-freedom model of the train and the artificial noise.The acceleration value is filtered by the improved Kalman filter,the low-pass filter and the ordinary Kalman filter.The results show that the maximum root mean square error of improved Kalman filter is 96%and 93%less than that of Butterworth low-pass filter and ordinary Kalman filter,respectively.The experiment on the train demonstrates that the device can accurately measure the acceleration value and calculate comfort and stationarity with better portability.