【作者】 张勇；

【导师】 唐功友；

【作者基本信息】 中国海洋大学 ， 海洋信息探测与处理， 2008， 博士

【摘要】 本论文针对时滞大于一个采样周期的网络控制系统,考虑NCS中存在的网络诱导时滞、数据包丢失和外部扰动等主要问题,以控制网络的服务质量为基础,围绕着NCS的性能指标,研究了NCS的建模、稳定性分析、最优控制、保成本控制、H_∞控制等问题,主要研究成果如下:1.针对具有网络诱导大时滞的网络控制系统,通过模糊融合技术对系统进行建模,将具有控制大时滞的系统转化为控制不含时滞的系统。利用时滞出现的概率作为隶属度,构成全局T-S模糊模型。基于模型设计了一种状态观测器,将计算扰动抵消,并对闭环系统进行了稳定性分析,给出了系统渐近稳定的条件。2.针对网络控制系统中网络诱导时滞可能大于一个采样周期的情况,通过模糊融合技术对系统进行建模,将具有控制时滞的系统转化为控制不含时滞的系统。运用Kalman滤波器进行白噪声最优状态估计,解决了最优扰动抑制控制律的物理实现问题,然后给出了前馈-反馈最优控制律的存在唯一性条件,并提出了控制律的具体设计算法。3.针对网络控制系统中同时具有数据包丢失和网络诱导时滞的情况,通过Try-Once-Discarded(TOD)技术对系统进行建模,将随机时滞对系统的影响转化为未知有界不确定项。在网络诱导时滞可能大于一个采样周期的情况下,将网络控制系统建模为具有事件率约束的异步动态系统(ADSs)。利用LMI方法的相关结论,给出网络控制系统指数稳定的充分条件以及数据传输成功率应满足的范围。4.研究受外部持续扰动的一类不确定性非线性网络控制系统的扰动抑制问题。提出了一种状态变量代换,将控制时滞转移到闭环控制回路之外,从而消除了时滞部分对控制系统稳定性的影响。利用内模原理给出了系统的无静差扰动抑制补偿器设计方法。运用Lyapunov稳定性理论和线性矩阵不等式技术,证明了保成本控制律的存在条件,并给出了无静差保成本控制器的设计方法。5.研究一类具有控制器增益摄动的大时滞不确定非线性网络控制系统的非脆弱保成本控制问题。首先将具有随机大时滞的网络控制系统模型化为具有不确定系数的时滞离散时间系统模型。然后利用Lyapunov稳定性理论和线性矩阵不等式方法,设计状态反馈非脆弱保成本控制器,使得闭环系统渐近稳定,并且系统的性能指标不超过某个确定的上界。同时给出非脆弱保成本控制律的存在条件和非脆弱保成本控制器的设计方法。6.研究一类大时滞不确定非线性网络控制系统的H∞保成本控制问题。首先将具有随机大时滞的网络控制系统模型化为具有不确定系数的时滞离散时间系统模型。然后利用Lyapunov稳定性理论和线性矩阵不等式方法,给出H∞保成本控制律的存在条件和H∞保成本控制器的设计方法。进而通过求解凸优化问题得到最优鲁棒H∞保成本控制器。最后部分总结了论文的主要内容,并对今后进一步的研究工作进行了展望。更多还原

【Abstract】 In this dissertation, the main problems of network-induced delay, data packet dropout and additive disturbance in networked control systems with long time-delay are considered. Basing on the quality of service of control networks, surrounding the quality of performance of NCS, the problems of modeling, stability analysis, optimal control, guaranteed cost control, H-infinity control for NCS are studied. The main contributions of this dissertation are as follows:1. Based on the fuzzy fusion technology, a modeling approach for networked control systems with network-induced long time-delay is proposed. The system with long control time-delay is transformed into the one without time-delay. By using the probability of time-delay appearance as membership degree, the global T-S fuzzy model is set up. According to the model, a states observer is designed and the disturbance arising from calculation is cancelled. Then the stability of the closed-loop system is analyzed and the conditions of asymptotic stability are proposed.2. Based on the fuzzy fusion technology, a modeling approach for networked control systems with network-induced time-delay which is probably longer than a sampling period is studied. The system with control time-delay is transformed into the one without time-delay. The Kalman filter is employed for the optimal state estimate of white-noise processes and the physically realizable problem of optimal disturbance rejection control law is solved. Then the existence and uniqueness conditions of feedforward and feedback optimal control law are proposed, and the algorithm of solving the optimal control problem is presented.3. By means of Try-Once-Discarded (TOD) technology, an approach of modeling for networked control systems with both data packet dropout and transfer delay is presented. The effect of random time-delay to the systems is regarded as the unknown bounded uncertain item. Based on the network-induced delay which is likely longer than a sampling period, the networked control system is modeled as an asynchronous dynamical system with rate constraints on events. Via the conclusion of the linear matrix inequality (LMI), the sufficient condition to the exponential stability of the networked control system is derived and the rage that must be satisfied by the data transmission rate is proposed.4. A disturbance rejection problem for a class of uncertain nonlinear networked control systems affected by additive persistent disturbances is considered. We present a state variable substitution which transfers the control time-delay to the outside of control closed-loop such that the impact of time-delay part to control system stability is eliminated. An approach to design zero steady-state error disturbance rejection compensator is proposed via the internal model principle. The Lyapunov stability theory and linear matrix inequality technology are employed to testify the existence conditions of guaranteed cost control law and to design a zero steady-state error guaranteed cost control strategy.5. A non-fragile control problem for a class of uncertain nonlinear networked control systems with long time-delay and controller gain perturbations is considered. Firstly, the NCS model with random long time-delay is transformed into a discrete-time system model with uncertain parameters. Then, the Lyapunov stability theory and the linear matrix inequality approach are applied to design a non-fragile controller, which results in that the closed-loop system is asymptotically stable and the system’s cost function value is less than a determinate upper bound. At the same time, the existence condition and the design approach of non-fragile controller are presented.6. An H-infinity guaranteed cost control problem for a class of uncertain nonlinear networked control systems with long time-delay is considered. Firstly, the networked control system model with random long time-delay is transformed into a discrete-time system model with uncertain parameters. Then, the Lyapunov stability theory and the linear matrix inequality approach are applied to present the existence condition and the design approach of H-infinity guaranteed cost controller. Furthermore, the optimal robust H-infinity guaranteed cost controller can be obtained through solving the corresponding convex optimization problem.Finally, the conclusions are given, and a proposition is indicated on the research work in the future.更多还原