학술논문
基于序贯自适应误差状态Kalman滤波的MARG姿态估计算法 / MARG attitude estimation algorithm based on sequential adaptive error state Kalman filter
Document Type
Academic Journal
Author
林新华; 赵陶然; 林威; 孙文涛; 江志健; 王梦蕾; LIN Xinhua; ZHAO Taoran; LIN Wei; SUN Wentao; JIANG Zhijian; WANG Menglei
Source
中国惯性技术学报 / Journal of Chinese Inertial Technology. 31(12):1175-1180
Subject
Language
Chinese
ISSN
1005-6734
Abstract
在由磁传感器、加速度计和陀螺仪(MARG)组合的航姿参考系统中,加速度计和磁传感器易受线性加速度和软硬磁干扰影响,从而降低系统姿态估计精度.针对上述问题,提出一种基于欧拉角误差的序贯自适应误差状态Kalman滤波算法,用于抗干扰的MARG姿态估计.通过建立基于欧拉角误差的线性系统误差模型,避免了基于四元数或欧拉角的MARG姿态估计算法中存在的非线性问题;同时,采用序贯Sage-Husa自适应算法,通过欧拉角误差量测值实时估计量测噪声参数,抑制线性加速度和软硬磁干扰对倾角和航向角估计精度的影响.为验证所提算法有效性,进行了抗干扰实验.实验结果表明,相比无序贯自适应误差状态 Kalman 滤波算法,所提算法的航向角估计最大误差降低了94%以上,倾角估计最大误差降低了 20%以上,提高了系统姿态估计精度,同时具有强鲁棒性.
In the attitude and heading reference systems based on magnetometer,accelerometer and rate gyro(MARG),the accelerometer and magnetic sensor are susceptible respectively to linear acceleration and soft and hard magnetic interference,which can compromise the accuracy of attitude estimation.To solve the problem,a sequential adaptive error state Kalman filtering algorithm based on Euler-angle error is proposed for anti-interference MARG attitude estimation.By establishing a linear system error model based on Euler-angle errors,the nonlinear problem of MARG attitude estimation algorithms based on quaternion or Euler Angle is avoided.At the same time,the sequential Sage-Husa adaptive algorithm is used to estimate the measurement noise parameters in real time by Euler-angle error measurements,and the effects of linear acceleration and soft and hard magnetic interference on the estimation accuracy of inclination and yaw are suppressed.In order to verify the effectiveness of the proposed algorithm,anti-interference experiments are designed.The experimental results show that compared with the non-sequential adaptive error-state Kalman filtering algorithm,the maximum error of the yaw estimation of the proposed algorithm is reduced by more than 94%,and the maximum error of the inclination estimation is reduced by more than 20%,which improves the attitude estimation accuracy of the system and has strong robustness.
In the attitude and heading reference systems based on magnetometer,accelerometer and rate gyro(MARG),the accelerometer and magnetic sensor are susceptible respectively to linear acceleration and soft and hard magnetic interference,which can compromise the accuracy of attitude estimation.To solve the problem,a sequential adaptive error state Kalman filtering algorithm based on Euler-angle error is proposed for anti-interference MARG attitude estimation.By establishing a linear system error model based on Euler-angle errors,the nonlinear problem of MARG attitude estimation algorithms based on quaternion or Euler Angle is avoided.At the same time,the sequential Sage-Husa adaptive algorithm is used to estimate the measurement noise parameters in real time by Euler-angle error measurements,and the effects of linear acceleration and soft and hard magnetic interference on the estimation accuracy of inclination and yaw are suppressed.In order to verify the effectiveness of the proposed algorithm,anti-interference experiments are designed.The experimental results show that compared with the non-sequential adaptive error-state Kalman filtering algorithm,the maximum error of the yaw estimation of the proposed algorithm is reduced by more than 94%,and the maximum error of the inclination estimation is reduced by more than 20%,which improves the attitude estimation accuracy of the system and has strong robustness.