基于联邦自适应滤波的分布式传递对准方法

Distributed Transfer Alignment Method Based on Federated Adaptive Filter

关键词:  分布式传递对准  联邦自适应滤波  卡尔曼滤波  机翼变形

Key words:  Distributed transfer alignment  Federated adaptive filter  Kalman filter  Wing deformation

摘要:

为解决复杂飞行环境下由于机翼的挠曲变形等因素导致分布式传递对准的精度下降和稳定性差的问题，提出了一种基于27维状态变量传递对准模型的新型联邦自适应滤波算法。首先，子滤波器采用基于R和P自适应更新的卡尔曼滤波算法并结合27维传递对准模型进行单点传递对准。然后，主滤波器采用基于误差协方差矩阵的原则，对多个子滤波器的冗余信息和信息权重进行信息融合以及自适应分配。基于实际飞行数据的半物理仿真实验表明：该方法在一定程度上提高了分布式传递对准的稳定性和精度，姿态估计稳定性得到明显提高，动态杆臂以及航向角估计精度分别提高了61.54%和42.35%。

Abstract:

In order to solve the problem of accuracy and stability degradation of distributed transfer alignment caused by wing deflection and deformation in the complex flight environment, a novel federal adaptive filtering algorithm based on the 27-dimensional transfer alignment model is proposed. Firstly, the transfer alignment based on the 27-dimensional transfer alignment model is carried out in the sub-filter, which adopt， the Kalman filter algorithm based on R and P adaptive update. Then, the redundant information and information weights of multiple sub-filters are fused and adaptively allocated in the main filter. The semi-physical simulation experiments based on actual flight data show that this method improves the stability and accuracy of distributed transfer alignment.The stability of attitude estimation is significantly improved. The level arm estimation accuracy and heading estimation accuracy are improved by 61.54% and 42.35%, respectively.