多尺度多传感器融合算法在微机电陀螺数据处理中的应用

The Application of Multi-Scale Sensor Fusion Algorithm to MEMS Gyroscope Data Processing

微机电陀螺以其成本低廉、安装方便而受到了人们的广泛重视,但人们期望其零偏稳定性能够进一步提高。多尺度多传感器融合算法是对陀螺仪输出的信号首先进行Kalman滤波;然后将多个陀螺的输出数据进行小波变换,在对应的尺度上将陀螺仪输出数据的小波变换系数进行数据融合,得到组合陀螺仪的各个尺度的小波变换系数;然后通过逆小波变换得到组合陀螺仪的输出信号。实验结果表明:经过将3个陀螺仪测量结果进行多尺度数据融合后的零偏稳定性比单个陀螺仪提高了一个量级,有效提高了微机电陀螺仪的精度。

The people pay more attention to MEMS gyroscope for its low cost and convenience. However, its zero-bias stability is still under development. It is steps of the multi-scale multisensor fusion algorithm that output signals of the gyroscopes are processed by the Kaman filtering; wavelet transforms of output data of the gyroscopes are performed, wavelet transform coefficients of output data are fused on the corresponding scale to obtain the wavelet transform coefficients of an integrated gyroscope; output signals of the integrated gyroscope can be reconstructed by using inverse wavelet transform. Fused results show that after doing multi-scale data fusion on the measurement result of N gyroscopes, the zero-bias stability improves by one quantity level.