基于动态事件触发时滞Lurie系统的变时滞补偿控制研究

Being derived from the engineering practices, Lurie system is an important theoretic model describing how the nonlinear uncertainties reflect the dynamic behaviors in a system. Meanwhile, time-delay always exists in many practical systems, which results in performance reduction or even disability. Thus the research on time-delay Lurie systems can present theoretical significance and application prospects. Through designing novel dynamic event-triggered schemes and combining time-varying delay compensation technique, some problems on control of time-delay Lurie systems will be deeply investigated in this project. Firstly, As for the issue on master-slave synchronization control, before the continuous-time controller and sampling one is studied respectively, based on input time-varying delay accurate or not, the controller design is accomplished based on output observer-based predictor, which guarantees synchronization error system to be absolute stable. The more general cases such as noise disturbance, discrete-time, and saturation control are also further considered. Now under the networked circumstance, as for centralized and decentralized sampling, the predictive controller will be respectively designed to solve the inaccurate communication time-varying delay compensation. Furthermore, the matrix decomposition technique will be utilized to establish the easy-to-test co-design, and the effect of quantization and packet loss are individually studied. By employing intermittent communication, the coordination control of multi-agent systems of Lurie type will be discussed. As for uniform communication time-varying delay or not, some novel distributed predictive control protocols will be proposed to obtain the corresponding event-triggered schemes. During the whole research, the characteristics of Lurie system will be fully employed to construct better Lyapunov functional and analytic method, which can greatly help extend the application areas of nonlinear controlled systems with multiple time-varying delays. Finally, the experiment on quadrotors’ flight control will be carried out to illustrate the effectiveness of the results on coordination control.

Lurie系统来源工程实际，是研究非线性不确定影响系统动力学行为的重要模型，而普遍存在的时滞现象会导致系统性能下降或不稳定，因此研究时滞Lurie系统具有重要理论意义和应用前景。项目结合新型事件触发机制与变时滞补偿技术，探讨时滞Lurie系统若干控制问题。针对主从同步控制，分别考虑连续时间与采样数据，当输入变时滞精确与否，完成基于输出观测预测器的控制器设计，使得同步误差系统绝对稳定，并拓展至随机干扰、离散时间及饱和控制。进而考虑网络控制下集中与分散式采样，设计通讯时滞未知时预测控制器，并利用矩阵分解法建立易验证的综合设计方案，并考虑量化丢包。研究间歇式通讯下Lurie型多智能体协同控制，针对通讯时滞相同与否，设计新型分布式预测控制协议，以获得相应事件触发机制。研究中充分利用系统特性改善泛函构造与解析方式，从而扩展多变时滞控制系统适用范围。最后，开展多四旋翼飞行实验，验证协同控制方法有效性。

非线性系统和网络控制系统等在很多领域具有很强应用背景，同时普遍存在的时滞会导致系统性能下降或不稳定。为了完成控制目标并降低控制代价，项目针对若干时滞控制系统，探索在新型事件触发与泛函构造等情况下，建立事件触发和控制器的综合设计。研究成果包括：1）按照原有研究计划，首先，针对中立型Lurie系统主从同步控制，提出拓展型交互凸组合技术，建立保守性较小的控制器设计方案；其次，改进现有的动态泛函构造，研究中立型时滞系统混合时滞相关下稳定性并给出保守性较小稳定性准则，同时利用该泛函构造技术，研究具有状态和输入时滞非线性系统抗饱和控制，基于静态输出反馈建立保守性较小的控制设计方案，最后，针对通讯时滞下二阶多智能体系统，建立分布式自适应事件触发下包容控制器设计；2）针对时滞中立型神经网络，利用分散式自适应事件触发和混杂时滞依赖技术研究同步控制，基于LMI技术建立相应事件触发和同步控制的综合设计方案；3）考虑通讯时滞下网络控制系统，提出新型自适应事件触发技术并综合考虑网络攻击和系统故障等影响，建立事件触发、观测器和控制器的综合设计方案；4）项目执行期为一年，共发表SCI论文8篇并申请2项专利。