针对一类网络控制系统,其网络诱导时延大于采样周期且存在数据丢包,研究了其稳定性和控制器的设计.在控制器端设计了一个状态预测器,其预测状态根据传感器的数据进行更新,然后用来确定控制作用.分析了预测误差的收敛性,证明了分离原理是成立的,并依据分离原理分别设计了状态预测器和控制器,给出了参数化设计方法.仿真例子验证了方法的有效性.
We study the stability and the design of a class of networked-control systems with long time-delay and datapacket-dropout. A state predictor is introduced in association with the controller; the predicted states are updated by the sensor data, and then are used to determine the control action. The convergence of the prediction-error is analyzed, and the separation principle is proved. By the separation principle, the parameters of the state predictor and the parameters of the controller are designed separately. Simulation results validate the effectiveness of the proposed methods.