Distributed Control of Constrained Compartmental Systems, with Applications to Large-Scale Infrastructures

该项目的研究目标是为线性隔室网络系统开发新的调节方法，这些方法可以直接和联合解决通常施加在这些系统上的物理和通信约束。分区系统可以被认为是一个相互连接的水库的集合，根据全局守恒定律交换某些物质的流量。因为这种物质代表一个物理量，如车辆的数量或流体的质量，这些调节方法必须能够产生所需的闭环行为，而不会产生要求这些变量变为负值的命令。此外，各种储存器之间的通信通常是有限的，并且因此必须仅基于本地测量来计算命令。所提出的方法建立在工具的鲁棒性和非线性控制理论，并利用特定的结构的隔间系统，使设计问题易于处理的大问题，同时将随时间变化的参数变化，随机扰动和不确定性。特别的努力和注意力将集中在这些方法的应用，以设定点调节，输出跟踪，和干扰拒绝在最近提出的国家空域系统和大规模的，重力驱动的，灌溉网络的空中交通流量的隔间模型。由于广泛的过程，可以建模为隔间系统，我们希望从拟议的研究产生的工具，以有利于许多重要的技术和社会应用，除了灌溉网络和空中交通流量控制。例子包括化学生产过程和一些分布式计算场景。技术成果将通过与美国宇航局艾姆斯研究中心的密切合作在空中交通管制界传播。本科生将积极参与灌溉委员会灌溉模拟及其验证试验台的建设。该试验台还将用作教学设备，作为当地高中学生关于节水挑战及其可能的工程解决方案的课程的一部分。

The research objective of this project is to develop novel regulation methods for linear compartmental networked systems, which can directly and jointly address the physical and communication constraints typically imposed on these systems. Compartmental systems can be thought of as a collection of interconnected reservoirs exchanging flows of some material according to global conservation laws.Because this material represents a physical quantity such as a number of vehicles or a mass of fluid these regulation methods must be able to generate the desired closed-loop behavior without producing commands requiring these variables to become negative. In addition, communication between various reservoirs is often limited, and the commands must thus be computed based on local measurements only. The proposed approach builds on tools from the theory of robust and nonlinear control, and exploits the particular structure of compartmental systems to make design problems tractable for large problems, while incorporating time-varying parameter variations, random perturbations, and uncertainties. Particular effort and attention will be focused on the application of these methods to set-point regulation, output tracking, and disturbance rejection in compartmental models recently proposed for air traffic flow in the National Airspace System and large-scale, gravity driven, irrigation networks. Due to the wide range of processes that can be modeled as compartmental systems, we expect the tools resulting from the proposed research to benefit many important technical and societal applications in addition to irrigation networks and air traffic flow control. Examples include chemical production processes and some distributed computing scenarios. The technical results will be disseminated within the air traffic control community through close collaboration with the NASA Ames research center. Undergraduatestudents will be actively involved in the construction of an Irrigation Board testbed for irrigation simulation and its validation. This testbed will also be used as a teaching device, as part of a course for local high school students on the challenges of waterconservation and its possible engineering solutions.