论文部分内容阅读
在红外成像跟踪系统中,通常仅能测量目标的角度信息,不能直接测量目标与观测站间的距离。研究了基于红外成像系统的被动测距技术,首先利用状态空间模型的分析方法建立被动测距的状态估计和参数学习的混合估计模型,然后介绍EM的基本原理和参数的最大似然估计。EM算法的E步利用粒子滤波和粒子平滑器来完成,实现被动测距的状态估计;M步利用梯度搜索的方法来求解参数。被动测距是一个带有未知参数的非线性系统的状态估计,文中利用状态估计与参数学习的状态空间模型来描述,并利用EM法来求解,为被动测距的求解提供了一条新的途径。模拟实验表明,基于粒子滤波和梯度搜索的EM方法能同时完成被动测距的状态估计和参数学习。
In the infrared imaging tracking system, usually can only measure the target’s angle information, can not directly measure the distance between the target and the observatory. The passive ranging technology based on infrared imaging system is studied. Firstly, the state space model is used to establish a hybrid estimation model of the state estimation of passive ranging and parameter learning. Then, the basic principle of EM and the maximum likelihood estimation of parameters are introduced. Step E of EM algorithm is implemented by using particle filter and particle smoother to realize the state estimation of passive ranging. Step M uses gradient search method to solve the parameter. Passive ranging is a state estimation of a nonlinear system with unknown parameters. It is described using the state space model of state estimation and parameter learning, and solved by the EM method, providing a new way for solving passive ranging . The simulation results show that the EM method based on particle filter and gradient search can accomplish the state estimation and parameter learning of passive ranging simultaneously.