卡尔曼滤波算法在AHRS姿态角解算中的应用

Kalman Filtering Algorithm to Calculating Heading and Attitude of AHRS System

AHRS航姿参考系统通过融合陀螺仪、磁强计、加速度计等MEMS传感器的数据来进行姿态角的解算。MEMS传感器精度较低,在受到外界干扰的情况下,检测到的原始数据会出现较为严重的失真现象,如果不对原始数据进行滤波处理,最后经过解算得到的姿态角会严重偏离真实值。常见的滤波方法有均值滤波、卡尔曼滤波、小波滤波等。当原始数据出现较大的波动的时候,均值滤波的滤波效果较差。实际应用中小波滤波算法复杂,适用性差。采用基于时间序列建模的方法对MEMS传感器原始数据建立数学模型,然后根据状态方程和测量方程对数据进行卡尔曼滤波,最后利用融合算法解算姿态角。实验证明,这种方法能够大大提高姿态角的精度。

AHRS system calculated heading and attitude by the method of fusing data from three MEMS sensors,such as gyroscope,magnetometer and accelerometer.However,due to low precision MEMS sensors,data gathered from sensors would deviate from true data when it is subject to some external disturbances,as a result,there would be a big error in the calculated heading and attitude if no action was taken.Common filtering algorithms included average filtering,kalman filtering and wavelet filtering.Average filtering behaved bad on filtering effect,especially when the original data fluctuated a lot.Wavelet filtering is not easy to be applicable because of its complexity.