Application of Transformed Cubature Quadrature Information Filtering in Distributed POS
Application of Transformed Cubature Quadrature Information Filtering in Distributed POS
复制标题
变换体积求积信息过滤在分布式POS中的应用
DOI：
10.1109/jsen.2021.3105403
复制
发表时间：
2021-10
影响因子：
4.3
通讯作者：
刘艳红
中科院分区：
文献类型：
--
作者：
王博;叶文;刘艳红
对于航空对地观测系统，机载分布式定位定向系统（POS）提供的运动参数的准确性对观测载荷成像质量至关重要。根据工作原理，分布式POS从机系统依靠主节点信息实现误差估计和修正，进而实现多节点高精度测量。由于误差模型的非线性，在估计过程中采用了非线性滤波算法。然后，为了获得更高的估计性能，一个变换体积正交信息滤波算法的分布式POS估计。该方法的主要贡献在于：（1）利用球面径向容积法则和Gauss-Laguerre求积法则生成sigma点;（2）对采样点进行正交变换，处理非局部采样;（3）采用信息滤波框架简化滤波更新过程，并将该算法应用于分布式POS非线性传递对准估计。通过飞行试验验证，该算法能有效提高分布式POS估计精度，包括姿态、速度、位置精度，可用于载荷成像质量的改善。
For aerial earth observation system, the accuracy of motion parameters provided by airborne distributed position and orientation system (POS) is crucial to the observation load imaging quality. According to the working principle, slave systems of distributed POS rely on the master node information to realize error estimation and correction, and then to achieve multiple node high accuracy measurement. Due to the nonlinearity of the error model, nonlinear filtering algorithm is used to the estimation process. Then in order to achieve higher estimation performance, a transformed cubature quadrature information filtering algorithm is developed for distributed POS estimation. In this method, the contributions are located that: (1) spherical radial cubature rule and Gauss-Laguerre quadrature rule are used to generate sigma points; (2) orthogonal transformation is conducted on the sampling points to deal with nonlocal sampling; (3) information filtering framework is adopted to simplify filtering update process, then the algorithm is applied in distributed POS nonlinear transfer alignment estimation. By the flight test validation, the outcomes are found that the developed algorithm can enhance the distributed POS estimation accuracy effectively, including attitude, velocity, position accuracy, then can be used to loads imaging quality improvement.