考虑双通道随机时延的非线性网络控制系统方法研究

The research on networked control systems (NCSs) has been receiving increasing attention due to its distinctive advantages and features. Most of the existing results have been developed either for linear NCSs or nonlinear NCSs with a special type of nonlinearities. Such a type of special nonlinear NCSs can be dealt with essentially by following the methodology for linear NCSs. In this project, we target at a class of general nonlinear systems, which not only can find wider and more general applications, but will be of more theoretical merits yet more challenging. The main objectives and works include the following four aspects. (1) By considering two-channel random delays ubiquitously existing in practical NCSs, we will design new two-mode-dependent robust control schemes for general nonlinear NCSs. (2) For the first time, we propose to employ the semi-Markov chain model to more accurately model and characterize the stochastic features of network delays, and further, we will develop a new framework for the analysis and synthesis of the resulting semi-Markov nonlinear NCSs. (3) In order to address practical constraints, we will incorporate the two-channel delays, the state constraints, the inpur constraints, and the external disturbances and noises into the analysis, develop a new compensation mechanism, and design the feasible nonlinear model predictive control algorithms. (4) We will comprehensively test and verify the proposed theoretical results on two NCS experimental test-bed: network-based quad-rotor vehicle and network-based robot manipulator. This proposed project is a type of application-oriented foundamental and theoretical research, and will find wide applications in industrial automation industries in our country.

网络控制系统研究已成为控制领域的研究热点，但主要集中在线性被控对象，或特殊类型的非线性被控对象。本项目针对一般类非线性被控对象，研究存在随机网络时延环境下非线性控制系统的分析和设计问题。主要内容包括（1）考虑双通道随机时延并存，设计依赖于双通道随机时延的非线性鲁棒控制算法；（2）首次采用半马尔科夫模型来更准确的描述随机网络时延，系统化的建立这类复杂半马尔科夫非线性系统的理论分析框架，提出设计控制器的新方法；（3）为解决实际应用中的实际约束条件，针对双通道随机时延、系统状态约束、输入约束、以及外部干扰和噪声等，提出基于预测控制信号的时延补偿机制，设计新的非线性网络系统预测控制算法；（4）在四旋翼飞行器和多自由度机械臂网络控制实验平台上验证理论结果。本项目属于应用基础研究，预期将取得严密深刻的原创性理论结果，并在试验平台上检验其实用性。研究成果在我国工业自动化企业中有重要的应用前景。

随着通信、计算机网络等以及控制技术的快速发展，网络控制系统研究已经成为系统与控制领域的研究热点，在工业自动化企业中具有广泛的应用前景。网络控制技术精简了控制系统线路，使得信息资源得到共享，减少了检测和故障诊断的成本，易于系统维护，提高了系统运行效能以及可靠性和灵活性。考虑实际网络控制系统中必然存在的双通道随机时延、模型不确定性、外界干扰、系统噪声等，本项目开展了非线性控制系统的分析和设计问题研究。本项目的研究内容主要包括四个方面。首先，采用马尔可夫链刻画动态随机带宽受限衍生的网络控制系统在有限时间内控制性能的下降与失稳问题。其次，考虑双通道时延并存的环境下，设计依赖于双通道随机时延的非线性鲁棒控制算法。再次，设计新的非线性网络控制系统的模型预测控制算法，解决了实际应用中的双通道随机时延问题，系统状态以及输入约束问题，以及外部干扰和噪声等。最后，构建了基于Internet的网络控制试验系统，其中包括四旋翼无人机和多个双轮移动式机器人，用来进行仿真和试验研究，进而验证了所提出的理论结果。本项目属于应用基础研究，不仅取得严密深刻的原创性理论结果，并通过仿真以及试验平台检验所提出的理论结果的有效性和实用性。