空间Kalman滤波与空时Kalman滤波算法  

Spatial Kalman filtering and spatial-temporal Kalman Filtering algorithm

作　　者：鄢丹青 

机构地区：[1]中国民航局第二研究所民航空管工程技术研究所工程技术室,四川成都610041

出　　处：《兰州理工大学学报》2016年第2期153-157,共5页Journal of Lanzhou University of Technology

基　　金：国家自然科学基金(61139003)

摘　　要：经典的Kalman理论是以观测按时间延续分布的方式建立的.针对观测按空间延展分布的情形,基于时空对偶性建立空间Kalman滤波的方法,进而提出针对观测同时在时间、空间两个方面展开的情形的空-时Kalman滤波算法.这些算法可用于包括多传感器信息融合等在内的广泛领域.仿真表明,相比于在不计代价(成本)的情况下精度最高的集中式多传感器融合算法,空-时Kalman滤波不仅具有滤波精度与之相当的优点,更重要的是,由于在计算复杂度上占有更大的优势,使算法实时性和有效性更为提高,更适用对实时性有更高要求的情形.The classic Kalman theory is established in the form of time-continuously distributed observation.Aimed at the case of space-continuously distributed observation and based on spatial-temporal duality,spatial Kalman filtering method is established.Further,a spatial-temporal Kalman filtering is proposed,which is based on both the time and spatial observation.These algorithms can be used in widespread fields,including multi-sensor information fusion.The simulation shows that,compared with the most high-accuracy centralized multi-sensor fusion algorithm with no spare the space-time Kalman filtering will have not only a filtering accuracy comparable with the former but also have improved real-time ability and effectiveness of the algorithm due to its superiority in complexity of computation.Therefore,it will be suitable for the case,where the requirement of real-time ability is stricter.

关 键 词：空间Kalman滤波 空-时Kalman滤波 多传感器信息融合 计算复杂度 

分 类 号：O231.1[理学—运筹学与控制论]