UKF在基于修正罗德里格参数的飞行器姿态确定中的应用

UKF Applied to MRPs-based Spacecraft Attitude Estimation

UKF是一种新的非线性滤波方法,在状态的时间更新阶段直接使用非线性模型,不引入线性化误差,而且不必计算Jacobi矩阵,相对于扩展卡尔曼滤波(EKF)不仅能提高滤波精度,而且更容易实现。修正的罗德里格参数(MRPs)是一种飞行器姿态参数,相对于四元数姿态参数,用MRPs表示飞行器的姿态时,状态误差方差不会产生奇异性,并且能在一定程度上减小估计的计算量。本文针对MRPs表示的无陀螺飞行器姿态系统,利用UKF设计了姿态估计器,并通过仿真验证了算法的有效性。

UKF (Unscented Kalman Filer) is a new nonlinear filtering method, which directly uses nonlinear models in the time-update phase. As a result, no error caused by linerization is introduced. Moreover, it is unnecessary to calculate the Jacobi matrix. Compared with the EKF( Extended Kalman Filter), the UKF not only can improve the precision of filtering, but also can be realized easily. MRPs is a kind of parameter to represent spacecraft attitude. When the attitude is presented by MRPs, the singularity of the state error covariance can be eliminated, and the computation load can be reduced to some extent. In this paper, the UKF is used to design attitude estimator for gyroless aerocraft systems, in which the attitude is presented by MKPs. The efficiency of the method is demonstrated by the results of simulation.