基于混合滤波的可展天线多节点状态估计

Mixture filter based state estimation of deployable antenna joints

为保证卫星可展天线精确和可靠地展开,提出一种基于混合滤波的可展天线多节点估计方法。该方法通过核粒子滤波器实现多模态概率密度的递推估计,滤波过程中采用分量迭代寻优模式使粒子向各自分量最大概率密度方向移动;并对混合参数重更新,以适应粒子寻优过程中的位置和权值变化。进一步结合可展天线模型建立自校正重要性函数,提高估计的准确性。实验表明该方法适用于天线在展开过程中的多节点状态估计。

For reliable and precise deployment of deployable antenna,a mixture filter based approach of state estimation of multiple antenna joints was presented.The approach estimated multi-modal probability density by kernel particle filter recursively.It moved the particles to the high probability areas of component probability density by iterative optimization,and then updated the component weights again to adapt to the change of positions and weights of particles.Furthermore,a self-correction important function relied on the dynamic model of antenna was built.Physical experiments showed that the approach was suitable to the state estimation of the deployable antenna joints.