CAREER: Advancing Network Configuration and Runtime Adaptation Methods for Industrial Wireless Sensor-Actuator Networks

十年来，无线HART和ISA100等工业无线标准的实际部署证明了在工业环境中使用基于IEEE 802.15.4的无线传感器-执行器网络(WSAN)实现可靠和实时无线通信的可行性。尽管由于多年的研究，WSAN在大多数情况下都能令人满意地工作，但它们往往很难配置，因为配置WSAN是一个复杂的过程，涉及理论计算、模拟、现场测试等任务。为了支持需要高数据速率和移动平台的新业务，工业WSAN正在采用5G和LORA等无线技术，并变得越来越分层、异质和复杂，这大大增加了网络配置的难度。这个职业项目旨在推进工业WSAN的网络配置和运行时适配方法。该项目的研究成果将显著增强工业WSAN的弹性和敏捷性，减少人类参与网络管理，从而显著提高工业效率并显著降低运营成本。通过提供更先进的无线存储区域网络，该项目的研究成果将极大地促进流程工业中无线存储区域网络的安装，并实现广泛的基于无线的新应用，从而影响经济、安全和生活质量。该项目加强讲课和课程项目材料，支持课程开发，为本科生和研究生创造研究机会，并为K-12学生建立外展计划。传统方法在很大程度上依赖于经验和经验法则，涉及在几个现场试验期间对网络负载或动态进行粗粒度分析，与此不同，该项目开发了一种严格的方法，通过获取无线研究社区积累的宝贵资源(例如，理论模型和模拟方法)来利用先进的机器学习技术来配置和适应WSAN。该项目开发了利用无线模拟和深度学习的新方法，以将高级网络性能与低级网络配置相关联，并在运行时高效地调整网络，以满足工业应用指定的性能要求。本项目通过试验台实验、案例研究和真实世界验证，展示了配备这些新方法的无线存储区域网络的性能。该项目的研究成果不仅影响工业WSAN，还影响其他复杂的无线网络，因为该项目为新的网络配置和运行时适配策略创建了一个可复制的模板，促进了无线网络管理的最先进水平。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

A decade of real-world deployments of industrial wireless standards, such as WirelessHART and ISA100, has demonstrated the feasibility of using IEEE 802.15.4-based wireless sensor-actuator networks (WSANs) to achieve reliable and real-time wireless communication in industrial environments. Although WSANs work satisfactorily most of the time thanks to years of research, they are often difficult to configure as configuring a WSAN is a complex process, which involves theoretical computation, simulation, field testing, among other tasks. To support new services that require high data rates and mobile platforms, industrial WSANs are adopting wireless technologies such as 5G and LoRa and becoming increasingly hierarchical, heterogeneous, and complex, which significantly increases the network configuration difficulty. This CAREER project aims to advance network configuration and runtime adaptation methods for industrial WSANs. Research outcomes from this project will significantly enhance the resilience and agility of industrial WSANs and reduce human involvement in network management, leading to a significant improvement in industrial efficiency and a remarkable reduction of operating costs. By providing more advanced WSANs, the research outcomes from this project will significantly spur the installation of WSANs in process industries and enable a broad range of new wireless-based applications, which affects economics, security, and quality of life. This project enhances lectures and course project materials, supports curriculum developments, creates research opportunities for undergraduate and graduate students, and establishes outreach programs for K-12 students. Different from traditional methods that rely largely on experience and rules of thumb that involve a coarse-grained analysis of network load or dynamics during a few field trials, this project develops a rigorous methodology that leverages advanced machine learning techniques to configure and adapt WSANs by harvesting the valuable resources (e.g., theoretical models and simulation methods) accumulated by the wireless research community. This project develops new methods that leverage wireless simulations and deep learning to relate high-level network performance to low-level network configurations and efficiently adapt the network at runtime to satisfy the performance requirements specified by industrial applications. This project demonstrates the performance of WSANs that are equipped with those new methods through testbed experimentation, case study, and real-world validation. The research outcomes from this project affects not only industrial WSANs but other complex wireless networks as this project creates a replicable template for novel network configuration and runtime adaptation strategies that advance the state of the art of wireless network management.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.