CAREER: Taming Uncertainties in Reliable, Real-Time Messaging for Wireless Networked Sensing and Control

无线网络传感和控制系统中的消息传递必须可靠且实时。这样的设计挑战需要新的模型和分布式协议的消息传递，可以处理干扰本地和准确地，计算概率路径延迟的基本问题是NP难的。本项目以单跳传输调度和多跳时空数据流控制为研究重点，提出了两个主要的研究任务：1）基于物理比K干扰模型，研究在线模型实例化的控制理论方法，解决大干扰范围和各向异性、非对称无线通信的挑战; 2）开发一种轻量级的方法来计算概率路径延迟，然后是用于实时消息传递的多时标自适应框架。特别地，该干扰模型将协议模型的局部性与物理模型的高保真度相结合，从而弥合了分布式实现的适用性与启用的调度性能之间的差距。通过在与相应动态相同的时间尺度上控制网络操作，所提出的多时间尺度适应框架确保了长期最优性，同时解决了短期动态。此外，该项目还包括一个综合、多层次、多组成部分的教育计划。教育活动将提高公众认识，提高学生的保留率，并使代表性不足的群体更多地参与计算。

Message passing in wireless networked sensing and control systems must be reliable and in real-time. Such design challenges require new models and distributed protocols for messaging that can handle interference locally and accurately, given the basic problem of computing probabilistic path delays is NP-hard. Focusing on single-hop transmission scheduling and multi-hop spatio-temporal data flow control, this project makes novel contributions by proposing two major research tasks: 1) investigation of control-theoretic approaches to online model instantiation, based on the physical-ratio-K (PRK) interference model, and addressing the challenges of large interference range as well as anisotropic, asymmetric wireless communication; 2) developing a lightweight approach to computing probabilistic path delays followed by a multi-timescale adaptation framework for real-time messaging. In particular, the PRK interference model integrates protocol model's locality with physical model's high-fidelity, thus bridging the gap between the suitability for distributed implementation and the enabled scheduling performance. By controlling network operations at the same timescale as the corresponding dynamics, the proposed multi-timescale adaptation framework ensures long-term optimality while simultaneously addressing short-term dynamics. Additionally, this project includes an integrated, multi-level, multi-component education plan. The education activities will raise public awareness, improve student retention and participation of underrepresented groups in computing.