中文摘要：

针对在工程实践中发生的测量数据随机丢失情况,提出了一种应用于非线性系统的滤波方法,该方法将基于序贯重要性采样的粒子滤波器应用于非线性、非高斯系统状态的在线状态估计。首先将测量数据丢失描述成满足一定条件概率分布的二元开关序列;然后基于似然函数设计方法,设计出测量数据丢失时的粒子滤波器算法;最后用本文方法对倒立摆系统状态估计进行了仿真。仿真实验表明,测量数据丢失时的粒子滤波器算法是有效的。

英文摘要：

Aimed at the case that sensor data may be missing randomly in practice, a filtering approach was proposed for the nonlinear systems, which applies a particle filter based on sequential importance sampling to the on-line state estimation of non-Gauss and nonlinear systems. The missing sensor data were described as a binary switching sequence which satisfies a certain conditional probability distribution; a particle filter algorithm in the presence of missing sensor data was designed based on likelihood function; the state estimation of a upside-down pendulum system was simulated by the proposed approach. The simulated results show the effectiveness of the proposed algorithm.