网络控制系统的一种统一的Markov跳变模型

对于网络诱导延迟的上界大于或小于一个采样周期的两种不同情况的连续时间网络控制系统,采用离散化与增广状态空间方法,建立了统一的Markov链离散时间跳变模型.由于模型含有反映系统的网络诱导延迟的大小与计算精度的参数,因此它具有更广泛的适应性.应用离散时间线性跳变系统理论,分析了保证系统均方稳定的充要条件.在所论及的保证系统均方稳定的设计算法中,给出了一种求内点法初始可行解的方法.在延迟带有一定任意性的车载倒立摆的网络控制设计中,该改进算法得到了有效应用.

Uniform Markovian jump model of networked control systems

For the two different cases that the upper bound of network-induced delays is longer or less than one sampling period in a continuous-time networked control system (NCS), by discretizing the model of a continuous-time system and applying the augmented state-space method to the model, a uniform discrete-time jump model governed by Markov chains was presented. The parameters of the model reflected both the length of the delays and computation precision, so this model had more widely adaptability. The existing theory of a jump linear system was applied to analysis the necessary and sufficient conditions guaranteeing the system's stability. An interior-point algorithm was applied to design stabilizing controllers, and an approach to get the feasible initial solution of the interior-point algorithm was proposed. This improved algorithm stated above was applied effectively to the networked control and design of a cart and inverted pendulum with random delays.