执行器受非光滑特性和故障约束的非线性系统网络控制设计

Compared with conventional industrial control systems, networked control systems have much more distinct advantages in the aspects of device management, remote operation, safety performance, organizing cost, etc. However, it is rather difficult to construct the networked control scheme for physical systems due to the limitation of bandwidth and the effect of some objective factors such as nonsmooth input characteristics and actuator failures, etc. Aiming at these problems, the project will conduct a research on networked control design of nonlinear systems. To reduce the communication rate of control signal and meanwhile achieve the adaptive inverse compensation for dead zone and backlash, the construction of inverse model for hybrid dynamics between event triggering mechanism and actuator is considered. To develop a robust adaptive event-triggered control strategy capable of suppressing hysteresis effectively, a network filter is designed to overcome the conflict between the discreteness of networked control signal and the differentiability of control signal required in hysteresis model. A stability analysis approach based on piecewise Lyapunov function and a new type of adaptive fault-tolerant design are further investigated, in order to develop the networked control strategy which can handle the coupling of nonsmooth input characteristics and actuator failures. Incorporated with prescribed performance techniques, the cases of fast time-varying input characteristics and high-frequency actuator failures can also be addressed. Through the study of the project, a new and novel adaptive networked control methodology will be developed, which is of great theoretical value and practical significance to both construct high-performance networked control systems and realize industrial informatization and intellectualization.

与传统工业控制系统相比，网络控制系统在设备管理、远程操控、安全性能、组建成本等方面具有更为明显的优势。但由于受到带宽限制以及非光滑输入特性和执行器故障等客观因素影响，构建物理系统的网络化控制方案存在很大难度。本项目将针对这些问题开展非线性系统网络控制设计的研究。考虑事件触发机制与执行器之间混合动态的逆模型构建，实现降低控制信号通信率的同时对死区、间隙自适应逆补偿；设计网络滤波器克服网络控制信号离散性和磁滞模型要求控制信号可导性之间的矛盾，形成有效抑制磁滞的鲁棒自适应事件触发控制策略；研究分段李雅普诺夫函数稳定性分析方法与新型自适应容错设计，建立应对非光滑输入特性与执行器故障耦合的网络控制策略，并结合预定性能技术实现对快时变输入特性与高频率执行器故障进行处理。本项目研究将形成一套自适应网络控制新方法，对构建高性能网络控制系统，实现工业信息化与智能化具有重要理论价值和实际意义。

网络化控制系统在航空航天、高速铁路、电力系统、远程医疗、智能制造等领域具有广阔应用前景。本项目采用不确定非线性系统模型对网络化控制的实际对象进行描述，并针对非光滑输入特性、执行器故障、通信时延等问题开展了高性能控制器设计及稳定性分析方面的研究工作。具体包括：1）针对具有非光滑输入特性的不确定非线性系统建立了基于量化或事件触发策略的自适应控制框架，解决了网络带宽受限下非光滑动态系统的网络化控制问题；2）提出了基于分段Lyapunov函数的稳定性分析新理论，并用于建立执行器反复故障情况下闭环系统的收敛性与鲁棒性条件，形成一系列对执行器故障具有自愈合能力的网络化控制新算法，已验证算法对无限数目具有正常数时间间隔的执行器故障均有效； 3）针对高性能网络化控制任务建立了不确定非线性系统的高速高精度控制策略，包括间接自适应模糊控制方法与直接自适应模糊控制方法，实现了跟踪误差在预定时间范围内收敛到预设精度。综上所述，本项目已针对具有非光滑输入特性与执行器故障的不确定非线性系统建立网络控制新理论和新方法，降低控制信号对过大网络带宽的依赖性，解决了带宽受限下非线性系统的网络控制设计问题，增强了网络控制系统对非光滑输入特性、系统不确定性、时变参数、未知外部扰动等因素的鲁棒性能以及对执行器故障的自适应容错能力，为网络化控制技术在实际工程系统的应用提供理论支撑。在本项目支持下，共发表论文20篇，其中SCI期刊论文15篇，《IEEE Transactions》12篇，EI会议论文5篇；申请中国发明专利5项，均已获得受理；获得省级自然科学一等奖1项。项目取得成果已超出预定目标。