摘要
针对低成本MEMS器件组合的姿态检测系统在运动加速度干扰下姿态估计精度较差等问题,提出了一种基于旋转矩阵卡尔曼滤波器(KF)的姿态解算方法。为了克服四元数法观测方程为非线性的缺点,该方法以旋转矩阵部分元素建立状态方程,并对量测加速度采用状态反馈估计的运动加速度进行补偿,减小了外部加速度的干扰,然后通过构造水平观测向量降低了计算复杂度,并给出了量测噪声协方差的推导。最后设计了卡尔曼滤波器对量测信息实现融合。动静态测试表明,该方法消除了累计误差,与无迹卡尔曼滤波(UKF)相比,提高了在运动加速度干扰下的姿态估计精度。
In order to solve the problem of external acceleration interference from attitude measurement system with low cost MEMS sensors,an attitude estimate algorithm based on rotation matrix Kalman filter is proposed.The algorithm uses the rotation matrix elements as description to overcome the shortcoming of nonlinear measurement equation,and the measurement acceleration is compensated by using state feedback estimated external acceleration,then the observation vector is structured to reduce computation complexity and system noise covariance is deduced.Lastly,a Kalman filter is used to achieve the multi-sensor information fusion.Experiments prove that the algorithm can improve the estimation precision compared with unscented Kalman algorithm and eliminate the accumulated error.
出处
《测控技术》
CSCD
2016年第2期52-57,共6页
Measurement & Control Technology
关键词
姿态估计
卡尔曼滤波
非线性
运动加速度
构造向量
attitude estimate
Kalman filter
nonlinear
external acceleration
construct vector
