An approach to design of feedback systems with time delay
An approach to design of feedback systems with time delay
复制标题
一种时滞反馈系统的设计方法
DOI：
10.1109/iecon.2005.1569200
复制
发表时间：
2005
期刊：
影响因子：
--
通讯作者：
K. Ohnishi
中科院分区：
文献类型：
--
作者：
K. Natori;K. Ohnishi
由于其独特的特性，时滞至今仍是控制工程中的一个难点和未解决的问题。因此，目前提出了多种补偿方法。其中最著名的是史密斯预测器。它是一种预测方法，等效地消除了闭环控制的时滞。状态预测控制是Smith预测器在状态空间中的扩展。但是，需要注意的是，在时间延迟波动和不可预测的情况下，这些方法的性能会下降。本文提出了一种设计时滞反馈系统的新方法。它基于扰动观测器的时延补偿方法和网络扰动的概念。由于该方法不需要建立时滞模型，因此适用于具有波动和不可预测时滞的系统。此外，还考虑了干扰的抑制技术。在此基础上，提出了一种可以分别考虑暂态响应和稳态响应的新型控制器。仿真和实验结果验证了该方法的有效性。
Because of the distinctive characteristics, time delay is even now a difficult and unresolved problem in control engineering. Therefore many compensation methods of it have been proposed so far. The most famous one of them all is Smith predictor. It is a kind of prediction method and equivalently eliminates time delay from closed control loop. State predictive control is briefly an extension of Smith predictor in state space. However it should be noted that the performances of these methods deteriorate in case that time delay is fluctuant and unpredictable. In this paper, a new approach to design of feedback systems with time delay is proposed. It is based on innovative time delay compensation method by disturbance observer and the concept of network disturbance. Because this method does not need delay time model, it is applicable to the system with fluctuant and unpredictable time delay. Furthermore it also considers suppression technique of disturbance. Then new controller that can be designed regarding transient and steady-state response separately is proposed. The validity of proposed method is verified by simulation and experimental results.