摘要
针对软故障条件下系统估计精度和容错性问题,将矩阵摄动理论引入联邦滤波器,提出并证明了"全局污染上界定理"。在此基础上,提出一种引入权衡因子的自适应信息分配策略,能够根据具体应用场合灵活调整局部滤波精度和容错性,尤其适合于软故障情况下改善无故障子滤波器的抗污染能力。最后通过SINS/GPS/ADS组合导航仿真证明该方法的有效性和可行性。
To analyze the system accuracy and fault tolerance under ＂soft＂ failures, the matrix perturbation theory is introduced in federated filter, and a globally contaminate maximum theorem is put forward and proved. Based on which, an adaptive information-sharing approach with a balance coefficient is presented. In terms of the occasion, the approach can neatly adjust local filters＇ accuracy and fault tolerance; especially, it is well suited to provide major improvements in good local filters＇ ability of resisting contamination under ＂soft＂ failures. Finally simulation results on SINS/GPS/ADS integrated navigation show that the validity and feasibility of the method.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2007年第11期1940-1944,共5页
Systems Engineering and Electronics
基金
"十一五"惯性技术预研项目基金资助课题(51309060302)
关键词
组合导航
信息分配
矩阵摄动
软故障
联邦滤波器
integrated navigation
information-sharing
matrix perturbation
＂soft＂ failure
federated filter
