对一类具有随机时延和输入约束的网络控制系统,利用变采样周期的方法,将连续的被控对象离散化,从而将网络控制系统建模为部分转移概率未知的Markov跳变系统。基于预测控制的滚动优化原理,提出一种模型预测控制策略,通过线性矩阵不等式的方法,给出保证整个闭环系统随机渐近稳定充分条件。仿真算例说明所提方法的有效性。
For a class of networked control systems with random delay and input constraints, applying variable-period sampling method, the networked control systems was modeled as a Markov jump systems with partly unknown transition probabilities. Based on the receding horizon optimization principle, a model predictive control scheme was proposed. By using the linear matrix inequality method, sufficient conditions were presented which guaranteed that the closed-loop system was asymptotically stable. Finally, an example was presented to illustrate the effectiveness of the proposed method.