多传感器系统的延长周期全信息融合滤波  

Full information fusion filtering based on period prolongation of multi-sensor system

作　　者：胡洲[1] 王志胜[1,2] 甄子洋[1] 

机构地区：[1]南京航空航天大学自动化学院,江苏南京210016 [2]国防科技大学机电工程与自动化学院,湖南长沙410073

出　　处：《传感器与微系统》2010年第5期37-40,共4页Transducer and Microsystem Technologies

基　　金：国家自然科学基金资助项目(60874037);南京航空航天大学博士学位论文创新与创优基金资助项目(BCXJ08-06)

摘　　要：提出了基于系统状态预测信息、子系统状态预测信息及量测信息的多传感器系统全信息融合滤波方法。针对多传感器系统中子系统状态估计信息与系统状态预测信息之间存在相关性的问题,采用延长系统融合器融合周期的方法来弱化它们之间的相关性,再进行直接融合来获得系统状态估计。基于"GPS姿态敏感器+陀螺"模式组合的卫星姿态确定系统的仿真结果表明了延长周期全信息融合滤波方法的有效性。A full information fusion filtering method based on the predictive information of the system state, the predictive information of sub-systems states and measnremental imformation for the multi-sensor system is presented in this paper. In allusion to correlation problem between state estimation information of the sub-system and state predictive information of the system in the multi-sensor system, a period prolongation method is used to weaken this correlation, and then the system state estimation is obtained by the direct fusion of the information. The simulation results of satellite attitude determination system based on the combination of GPS and gyroscope show the effectiveness of the full information fusion filtering based on the period prolonga-tion.

关 键 词：多传感器系统 全信息融合 滤波 姿态确定 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置]