一种基于改进卡尔曼滤波的姿态解算算法

针对多传感器融合姿态解算精度不高的问题,本文提出一种改进的卡尔曼滤波算法,即高阶线性互补滤波与扩展卡尔曼滤波(Extended Kalman Filter,EKF)相结合的融合算法。该数据的融合是基于加速度计、陀螺仪传感器频率特性和姿态角的微分方程建立的系统模型,将互补滤波的姿态角数据作为该系统模型的观测值,利用EKF算法对加速度计、陀螺仪、磁力计进行数据融合。高阶的互补滤波和EKF的融合算法能够有效的解决陀螺方向的估计偏差,为了证明该算法的可行性,用搭载IMU(InertialSmeasurementSunit)模块的四旋翼飞行器进行了动态和静态的实验,分析对比了最新导航算法、经典卡滤波算法和该融合算法滤波的效果。实验结果表明：本文提出的高阶无源线性互补滤波和EKF相结合的融合算法,无论在静态还是动态的实时性情况下,都能很明显的去除噪声和抑制姿态角的漂移,且提高了姿态角的精度。

An attitude solving algorithm based on improved kalman filter

Aiming at the problem that the multi-sensor fusion attitude resolution is not high, this paper proposes an improved Kalman filter algorithm, which is a fusion algorithm combining high-order linear complementary filtering and Extended Kalman Filter (EKF). The fusion of the data is based on the accelerometer, the gyro sensor frequency characteristics and the differential equation of the attitude angle. The complementary filtered attitude angle data is used as the observation value of the system model, and the EKF algorithm is used for the accelerometer and the gyroscope. The magnetometer performs data fusion. The high-order complementary filtering and EKF fusion algorithm can effectively solve the estimation deviation of the gyro direction. In order to prove the feasibility of the algorithm, the dynamic and static experiments are carried out with a quadrotor equipped with an IMU module. The latest navigation algorithm, classic card filtering algorithm and the filtering effect of the fusion algorithm are compared. The experimental results show that the proposed high-order passive linear complementary filtering and EKF fusion algorithm can remove noise and suppress the attitude angle drift in both static and dynamic real-time situations. The accuracy of the attitude angle.