摘要
利用Kalman滤波器对MEMS陀螺随机漂移进行估计和补偿,需要将随机漂移的自回归滑动平均(ARMA)模型转化为相应的状态空间模型,从而有必要对模型之间的转换问题进行深入研究。针对国内外有关文献提出的三种转化形式,从数学角度进行了证明,并结合MEMS陀螺的试验数据,分别采用这三种状态空间模型进行了随机漂移和角速率估计试验。试验结果分析表明,采用状态空间模型1不能同时对陀螺随机漂移和角速率做出正确的估计,采用状态空间模型2、3的估计结果正确且滤波效果好、实时性强,更适用于对MEMS陀螺随机漂移的估计和补偿。
To estimate and compensate random drift of MEMS gyro effectively by using Kalman filter,the autoregressive moving average（ARMA）model of MEMS gyro needs to be transformed to the state space form correctively,and then it is necessary to further research on those transformations.Based on the 3 kinds of transformation from correlative references at home and abroad,and combined with the data from MEMS gyro,the test of random drift estimation with those 3 state space models is put forward.The test result indicates that the random drift cannot be estimated correctively by using the model 1;on the other hand,the model 2 and 3 are more valuable and suitable for the estimation and compensation of MEMS Gyro random drift for the merits of correct results,better filtering and timely performance.
出处
《传感技术学报》
CAS
CSCD
北大核心
2011年第6期853-858,共6页
Chinese Journal of Sensors and Actuators
基金
总装预研基金项目(9140A09031008CB01)
