一类线性离散时间结构随机跳变系统的逼近滤波算法  

An approximate optimal filtering for discrete stochastic systems with conditional Markov structure

作　　者：方洋旺[1] 王洪强[1] 伍友利[1] 

机构地区：[1]空军工程大学工程学院,陕西西安710038

出　　处：《控制理论与应用》2009年第8期889-892,共4页Control Theory & Applications

基　　金：国家自然科学基金资助项目(60674031);国家高科技发展计划"863"计划资助项目(2007AA701405);空军工程大学工程学院优秀博士学位论文创新基金资助项目(BC07004)

摘　　要：提出了具有条件马尔可夫跳变结构的离散时间随机系统的条件滤波方法,应用随机变结构系统的性质对滤波算法进行简化处理,并将后验概率密度函数用条件高斯函数来逼近,得到具有条件马尔可夫结构离散随机系统的逼近最优滤波算法,最后给出滤波算法的计算步骤并仿真验证了算法的正确性.A conditional filtering algorithm is presented for a discrete stochastic system with conditional Markov structure. This filtering algorithm is simplified by making use of the variable-structure character of the stochastic system; and the posterior probability density function is approximated by a Gaussian function. Thus, an optimal filtering for conditional Markov structure discrete stochastic is obtained. The computation process of this filtering algorithm is given; and its effectiveness is shown by simulation results.

关 键 词：随机系统 变结构系统 条件马尔可夫结构 逼近最优滤波 

分 类 号：TN713[电子电信—电路与系统]