介绍了前向线性预测滤波算法的基本原理,提出了一种自适应滤波过程中各参数的确定方法,对某硅微陀螺的静态

漂移信号和实际动态信号进行了处理,给出了静态漂移信号滤波前后的Allan 方差和标准差的大小,对滤波前后的误差大小

和误差分布进行了分析,并与小波中值滤波效果进行了比较。结果表明,前向线性预测滤波方法无论是在去噪效果,还是实

时性等方面,都明显优于小波中值滤波-Principle of t he forward linear prediction ( FL P) is int roduced. How to destine all parameter s of

t he FL P by an adaptive met hod is given. A silicon gyroscope static drif t and t he dynamic measurement is

processed by FL P. The Allan variance and t he standard variance is used to compare t he filtering result .

The error and it s dist ribution is listed. The FL P is compared with t he median2wavelet filtering. The result

shows t hat the FL P is superior to t he median2wavelet in the de2noise effect and real2time and so on.