摘要
在对适用于INS/GPS组合导航的H∞滤波方法进行了深入研究基础上,提出了一种改进的H∞多重渐消容错滤波算法。首先给出了H∞滤波算法,然后对渐消记忆滤波方法进行了有效的改进,提出了多重渐消因子矩阵对称加权的滤波算法和残差归一化的多重渐消滤波方法,有效避免了残差量级差别所导致的残差检测失效和修正作用失配,最后提出了一种双重故障检测的容错策略,构造了改进的鲁棒H∞容错滤波器。仿真结果表明,改进的H∞容错滤波方法较常规Kalman滤波对系统模型和噪声的不确定干扰具有很强的鲁棒性,在有色噪声情况下具有高于常规Kalman滤波的精度,对于故障具有较强的检测能力和较好的组合效果,更加贴近于工程应用需求。
On the basis of in-depth study for H∞ filtering in INS/GPS integrated navigation, an improved H∞ muhiple fading fault-tolerant filtering algorithm is proposed. First, an H∞ filtering algorithm is demonstrated, and then the fault-tolerant filtering method is improved effectively. The multiple fading filtering algorithm based on symmetrical proportion-weight of matrix and the algorithm based on the normalized residuals method are presented to avoid effectively invalidation detecting and correcting for residuals. A fault-tolerant strategy of double fault diagnosis is proposed and the improved robust H∞ tolerant filter is designed ultimately. The simulation results indicate that the proposed algorithm has better robustness and accuracy to model uncertainties and statistics characteristics of disturbance variations than common Kahnan filter. The improved filtering algorithm has a great detecting capability and combining performance, and it is fit for requirements of engineering.
出处
《宇航学报》
EI
CAS
CSCD
北大核心
2009年第3期930-936,共7页
Journal of Astronautics
基金
国家863课题资助项目(2008AA7050411)
国家自然科学基金(60874093)
