Agent-Based Approach to Smart Sensing for Health Monitoring of Civil Infrastructure

这项研究旨在利用先进的传感、微处理、数字信号处理、无线通信和损伤诊断方法，为创新的智能传感和健康监测概念奠定广泛的基础。这项工作旨在为极端事件（如强震）以及长期恶化提供近实时的结构状况评估。目标是开发一种计算框架，该框架可以充分利用连续和周期性无线监测策略，并且可以自动可靠地检测结构中可能的损坏位置或恶化区域。自主智能体和多智能体系统是分析、设计和实现以智能传感器为代表的复杂分布式计算系统的重要新途径。Mote平台由加州大学伯克利分校在国防高级研究计划局（DARPA）的资助下开发，首次为广泛的智能传感研究提供了一个开放的硬件/软件环境，并将被用作这项研究的基础。综合研究计划有四个主要任务：(i)为Berkeley-Mote平台设计和构建新的高灵敏度加速度和应变传感器，以促进民用基础设施系统的全球/本地监测；（ii）构思、开发和实施基于agent的框架，用于基于Berkeley-Mote平台的民用基础设施系统的监测和损伤检测；（iii）开发新的结构健康监测算法，可在智能传感器的分布式计算环境中执行。（iv）在实验室和现场环境中对硬件和软件进行原型和测试。这一努力包括与三个机构的重要美中合作部分。这项研究的成功完成预计将加速保护我们各自国家基础设施的创新战略的经济和实际实施。

This investigation seeks to develop a broad foundation for innovative smart sensing and health monitoring concepts utilizing advanced sensing, micro-processing, digital signal processing, wireless communication, and damage diagnostic methods. This work is aimed at providing near real-time structural condition assessment for extreme events (e.g., strong earthquakes), as well as long-term deterioration. The goal is to develop a computational framework that can take full advantage of the continuous and periodic wireless monitoring strategy and that can automatically and reliably detect possible damage locations or deteriorated regions in a structure. Autonomous agents and multi-agent systems represent an important new way of analyzing, designing, and implementing complex distributed computing systems represented by smart sensors. The Mote platform developed at the University of California at Berkeley with funding from the Defense Advanced Research Projects Agency (DARPA) offers, for the first time, an open hardware/ software environment for broad smart sensing research and will be employed as a basis for this research. The comprehensive research plan has four primary tasks: (i) design and construct new high sensitivity acceleration and strain sensors for the Berkeley-Mote platform to facilitate global/local monitoring of civil infrastructure systems, (ii) conceptualize, develop, and implement an agent-based framework for monitoring and damage detection in civil infrastructure systems based on the Berkeley-Mote platform, (iii) develop new structural health monitoring algorithms that can be executed in the smart sensor's distributed computing environment, and (iv) prototype and test developed hardware and software in both the laboratory and field environment. This effort includes a significant US-China collaborative portion with three institutions. Successful completion of this research is expected to accelerate economic and practical implementation of innovative strategies for protecting our respective nations' infrastructure.