基于MUBF算法的微机械陀螺输出降噪方法

Noise reduction of MEMS gyroscope based on MUBF algorithm

针对微机械(micro electro mechanical system,MEMS)陀螺输出漂移不确定性,提出采用最小上限滤波(minimum upper-bound filter,MUBF)算法实现MEMS陀螺输出信号降噪处理,该算法将漂移看作陀螺输出信号中的未知干扰,通过获取漂移变化方差上限,利用凸优化动态寻优得到角速率估计。相比卡尔曼滤波算法(Kalman filter,KF),MUBF算法可以在陀螺输出漂移模型未知的情况下工作,弱化陀螺信号降噪处理条件。陀螺静态和动态实验结果表明:MUBF算法能够有效降低陀螺噪声且优于KF算法降噪效果,该算法为MEMS陀螺降噪研究提供新思路。

Due to uncertain bias drift of the micro electro mechanical system （MEMS） gyroscope, the minimum upper-bound filter （MUBF） algorithm is presented to reduce the signal noise of MEMS gyroscope. Here, the bias drift is treated as unknown disturbance. In pursuit of the maximum variance value of the bias drift, the convex optimization process is then utilized to estimate the angular rate output of the MEMS gyroscope. Compared with the Kalman filter （KF） algorithm, the MUBF algorithm can work without a priori delicate bias drift evolvement model, while the condition of the filter existence is easily satisfied. Experimental results show that filtering effectiveness based on the MUBF algorithm is remarkable to suppress the bias drift of the MEMS gyroscope and better than that of the KF algorithm. The proposed method provides a new way to decrease the signal noise of the MEMS gyroscope in practice.