基于自适应滤波的MEMS姿态确定方法

Research on the Attitude Determination of MEMS Based on Adaptive Filter

针对消费类电子设备对姿态测量系统的需求,本文提出了一种基于MEMS加速度计、陀螺仪和磁强计的九轴姿态确定算法。针对实际系统中传感器量测噪声未知的情况,首先介绍了一种基于矢量观测器的矩阵Kalman滤波姿态确定算法,然后利用残差匹配技术,设计了一种基于残差匹配的自适应滤波方法。论文采用自适应滤波对传感器量测噪声进行估计,并将估计的量测噪声代入线性矩阵Kalman滤波算法,有效解决了线性矩阵Kalman滤波需要准确量测噪声统计信息的缺陷。最后设计了仿真实验验证本文提出的算法,并将其与线性矩阵Kalman滤波算法比较。仿真结果表明,自适应矩阵Kalman滤波的姿态旋转误差角为0.609 1°,标准差为0.300 9°,能够有效的估计传感器量测噪声,并具有更高的姿态确定精度和稳定性。

In view of the demand for attitude measurement system in the consumer electronics devices,a novel attitude determination method based on MEMS accelerometer,MEMS gyroscope and MEMS magnetometer is put forward in this paper. Because the statistics information of the measurement noise of the inertial sensors in the actual system is unknown,a matrix Kalman filtering for attitude determination algorithm based on the vector observer is introduced. In addition,the adaptive filtering method is designed on the basis of the residual matching technology.With this method,the measurement noise can be estimated,and the difficulty of obtaining the accurate statistics of measurement noise for the traditional matrix Kalman filtering is overcome. Finally,the simulation test is designed,and the results show that the rotation error angle is 0.609 1°,and the standard deviation is 0.300 9°,which indicates that the measurement noise of the sensors can be determinated effectively,and the adaptive filter is more accurate and more stable by comparing with the traditional matrix Kalman filtering.