An approximate optimal filtering for discrete stochastic systems with conditional Markov structure
一类线性离散时间结构随机跳变系统的逼近滤波算法

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.