网络系统的马尔科夫时滞预测控制

Markov delay predictive control for network systems

针对非线性网络控制系统存在网络诱导时延的问题,提出一种时滞补偿控制方法。首先建立非线性网络系统的T-S模糊模型,将网络诱导时延视为系统输入时滞;利用马尔科夫链对网络诱导时滞进行建模,在此基础上计算状态转移概率矩阵,依据最大转移概率法对下一步的时滞进行预测;采用遗传算法对状态反馈控制器进行寻优,得到T-S模糊模型每个子系统针对各个时滞的最优状态反馈控制律;通过时滞时间的隶属度函数融合得到非线性网络控制系统的模糊预测控制器;仿真算例验证该方法的可行性与有效性。

In view of network induced time delay in nonlinear networked control systems, a time delay compensation control theme was proposed. T-S fuzzy model was established for nonlinear network control systems firstly. Then the network induced delay was modeled as Markov chain, and delay was predicted. In the following, by using genetic algorithm, the optimal state feedback control law for each subsystem with each delay was established. And then fuzzy predictive controller of the nonlinear networked control systems was obtained through membership function. Simulation example shows the effectiveness of the proposed method.