可处理多个无序量测的前向滤波算法  

Forward-filtering algorithm for dealing multiple out-of-sequence measurements

作　　者：王炜[1,2] 黄心汉[2] 王敏[2] 王公宝[1] 

机构地区：[1]海军工程大学理学院,湖北武汉430033 [2]华中科技大学控制科学与工程系,湖北武汉430074

出　　处：《华中科技大学学报（自然科学版）》2012年第7期11-14,共4页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(60873032);海军工程大学校基金资助项目(HGDQNEQJJ11001)

摘　　要：基于等价量测和前向滤波法,推导出一种单个无序量测算法Fl1,并说明了该算法与Al1算法、FPFD算法以及AA1算法之间的等价性.以Fl1算法为基础提出可以处理同时到达融合中心的多个不同单步延迟无序量测的最优算法CZ1.所提新算法的性能不依赖过程噪声的离散化模式,且无须对状态转移矩阵求逆.仿真结果表明在过程噪声DCM下CZ1算法与CA1算法的性能是相同的.The equivalent measurement approach and forward-filtering method were used to derive an single out-of-sequence measurement processing algorithm denoted as Fl1.It was demonstrated that the algorithm was equivalent to algorithm Al1,FPFD（forward-prediction fusion and decorrelation） algorithm and algorithm AA1.Subsequently,based on algorithm Fl1,an optimal update algorithm with multiple one-step-lag out-of-sequence measurements,denoted as CZ1,was presented.The proposed algorithm did not require the inverse of the state transition matrix,and their performance was independent on the discretization form of process noise.The simulation results show that the CZ1algorithm has the same preformance as the CA1algorithm in discretized continuous-time model of the process noises.

关 键 词：信息融合 无序量测 等价量测 前向预测 滤波更新算法 

分 类 号：TP14[自动化与计算机技术—控制理论与控制工程]