中文摘要：

分析了递推最小二乘法的基本原理，提出采用渐消的最小二乘参数实时递推法对某压电陀螺的随机漂移进行实时自回归（AR）建模后，并进行卡尔曼（Kalman）滤波处理。同时，该文还提出利用一种自适应补偿法对陀螺随机漂移进行补偿，最终将陀螺精度提高了3倍多。仿真结果表明，无论是处理效果，还是硬件实现的实时性都表明，这些处理方法可实现并能得到应用。

英文摘要：

The principle of the RLS is analyzed. After the piezoelectric gyroscope random drift is AR modeled by the factor fading RLS, the Kalman filtering is used to de-noise the gyroscope random drift. A self-adaptive compensation way can improve the gyroscope accuracy more than 3 times. The simulation result shows that these algorithms of the piezoelectric gyroscope random drift can be implemented either in the real-time and the processing result.