基于极大似然的单传感器误差配准算法  

A Maximum Likelihood-Based Single Sensor Bias Registration Algorithm

作　　者：宋强[1] 何友[2] 熊伟[2] 

机构地区：[1]中国人民解放军91655部队 [2]海军航空工程学院信息融合技术研究所

出　　处：《宇航学报》2011年第8期1826-1832,共7页Journal of Astronautics

基　　金：国家自然科学基金(61032001;60801049)

摘　　要：对单传感器系统误差配准问题进行了研究,首先理论分析了传感器量测系统误差对固定目标探测定位的影响,给出了传感器运动将造成固定目标产生虚像做类似运动的结论;在此基础上,提出了一种基于极大似然的单传感器误差配准算法,该算法基于极大似然准则,利用运动传感器对固定目标虚像位置的多时刻跟踪滤波信息,构建对传感器量测系统误差的极大似然估计,并通过进一步迭代估计获得系统误差最终估计结果,从而能够利用位置未知的固定目标实现运动平台单传感器量测系统误差的精确配准。The single sensor bias registration problem is researched in this paper.First,the influence of sensor measured bias on a fixed target location is analyzed theoretically,and the conclusion that sensor movement will make the target do similar movement as a virtual target is given.Then,on the basis of analysis,a single sensor bias registration algorithm based on maximum likelihood is presented.Using tracking information of the virtual image in multiple times,the registration algorithm is used to generate the maximum likelihood estimation of sensor measured bias and to get the final estimation by an iterative process.Thereby,all of these show that the algorithm can be used to achieve an exact registration of single moving sensor by using a fixed target with unknown location.

关 键 词：系统误差 极大似然估计 误差配准 传感器组网 

分 类 号：TP953[自动化与计算机技术] TN957[电子电信—信号与信息处理]