基于虚拟并发数据流和有效控制数据驱动的无线网络化系统协同设计

Wireless networked control systems (WNCSs) have been a hot research topic in recent years. In the WNCSs field, three key issues still deserving further research are: 1) the wireless network bandwidth utilization is inefficient, 2) the conservatism of networked control methods should be further reduced, and 3) more effective scheduling and control co-design approach is desired. To tackle the above problems, the key idea behind this project is "bottom-up first and then top-down". Firstly, we utilize simulated multiple data flows to make fully use of the wireless bandwidth, and further reduce the conservatism of networked control methods by employing the effective control signals at the analysis and synthesis stage. Then, under this framework, the Network Pricing and Nash Equilibrium theories are used for the optimization and co-design of WNCSs. Based on the above idea, the problems of wireless network transmission rate control, the WNCSs modeling, the stability analysis, the controller design, and the co-design of scheduling and control, will be addressed in this project, where comprehensive study will be carried out on the experimental platforms. The obtained results of this project are expected to show better performance than the existing methods, and therefore could provide new research ideas and solutions for the research of WNCSs co-design.

目前无线网络化控制系统（WNCSs）已成为学术界研究的热点，但无线网络带宽利用率低、调度与控制协同设计程度不高以及控制算法保守性有待进一步降低三大问题依然突出。针对上述问题，本项目提出"先自底向上，后自顶向下"的研究思路：首先基于虚拟并发数据流充分饱和带宽的跨层速率传输控制算法来有效提高网络利用率，基于有效控制数据驱动和Delta算子来进一步降低控制算法保守性；在此基础上，进一步利用Nash均衡、网络定价和优化算法等技术进行调度和控制参数的协同优化设计。基于上述思路，本项目系统深入地研究无线网络的跨层速率传输控制算法、闭环系统建模、稳定性分析与控制器设计、调度与控制协同优化等问题，并在WNCSs实验平台上予以实现。其研究成果有望克服现有WNCSs分析设计方法的不足，大幅提高WNCSs系统的整体性能，从而为WNCSs的协同设计提供一种新的研究思路和解决方案。

无线网络化控制系统已成为学术界研究的热点，但无线网络带宽利用率低 、调度与控制协同设计程度不高以及控制算法保守性有待进一步降低三大问题依然突出。在国家自然基金的支持下，项目组系统地研究了网络控制系统建模、稳定性分析以及控制器设计等问题，圆满的完成了预定研究内容并取得了一系列研究成果。到目前为止，项目组获CAA自然科学二等奖一项，在IEEE Trans. on Fuzzy Systems、IEEE Trans. on Cybernetics等国际杂志上发表（含录用）SCI论文17篇，国际会议特邀报告报告2人次；报告已培养博士2名，硕士1名。