基于自适应联邦滤波的传递对准算法

Transfer alignment algorithm based on adaptive federated filter

为了提高传递对准的对准速度和对准精度,研究了基于自适应联邦滤波的"速度+姿态角"传递对准的算法,在状态变量中加入了安装误差角和挠曲变形角,在算法中进行了估计并补偿。为了保证滤波的实时性,采用联邦滤波的方法,分别建立了速度匹配子滤波器、姿态角匹配子滤波器和主滤波器的模型,给出了状态方程和量测方程。在子滤波器中利用模糊控制器对噪声特性进行了自适应调整以解决系统噪声和量测噪声是未知情况下滤波发散或者精度不高的缺点从而增强系统的鲁棒性。最后在载体匀速直线加三轴摇摆的运动模型下进行了仿真,结果表明该方法能够有效地估计安装误差角和部分挠曲变形角,并且能够以一定精度完成初始对准。

The ＂velocity＋attitude＂ transfer alignment algorithm based on adaptive federated filter was studied to improve the alignment speed and accuracy, which added the misalignment and the flexural distortion angle into the state variables and made estimation and compensation for them. The models of velocity match sub-filters, attitude match sub-filters and main filter were established based on federated Kalman filter for ensuring real-time filtering. The noise was adjusted adaptively in sub-filter using fuzzy control module to overcome the disadvantages of divergence and low accuracy. At last the simulation was made based on the motion model of ＂uniform linear motion ＋ triaxial rocking＂. The result shows that the method can effectively estimate the misalignment and the partial flexural distortion angle, and realize the transfer alignment with certain accuracy.