S&AS: FND: Cognitive and Reflective Monitoring Systems for Urban Environments

随着城市人口的增长，迫切需要能够使城市系统更加有效和高效的技术解决方案。许多设想的智慧城市系统，例如智能交通、车辆和安全网络，需要访问广泛的在线/离线数据，这些数据描述了整个城市的运行状态和发生的事件。然而，城市范围内的传感器和处理系统的实际部署面临着一些挑战，包括财务成本、在严格的服务质量要求下传输大量数据的网络资源的可用性，以及与使用相同资源的其他现有服务的共存。该项目的目标是开发一个移动传感器的认知和反射网络，能够利用渗透到表征城市环境的分层通信、处理和传感基础设施的智能概念，最大限度地减少其对关键城市通信资源的影响。该项目的成功实现将为在技术和成本限制的背景下实现实时城市物联网（IoT）系统的承诺做出重大贡献。该项目还包括一个多层次的教育、指导和推广计划，以培训下一代物联网系统设计师和专业人员。所提出的系统增强了单个传感器通过智能分层架构做出导航决策的能力，该架构能够从全局计算目标的角度提供有关其生成数据的有用性的实时反馈。然后，实时适应过程在架构的各层上表达，其中代理动态地学习效用模型并在不同的地理和时间尺度上进行控制。这些实用模型用于协调移动传感器在城市内的导航，以实现事件和动态过程的检测和监控。该项目的成果是第一个此类架构，其中认知和智能遍布城市传感、通信和处理基础设施的各个规模。此外，该项目设想的系统代表了边缘计算架构为数不多且最具创新性的示例之一，其中低延迟处理的可用性用于优化系统的操作。 为这种复杂的分层场景构建框架，涉及具有不同传感和计算能力的设备，提出了固有的技术挑战，这些挑战将通过在分布式和分层学习和机器人导航领域产生多项创新来解决。

As the urban population grows, a pressing need arises for technological solutions capable of making city systems more effective and efficient. Many of the envisioned Smart City systems, such as intelligent transportation, and vehicular and security networks, require having access to a wide spectrum of online/offline data that characterizes the state of operation and the events taking place throughout the city. However, the practical deployment of city-wide sensor and processing systems faces several challenges, including financial cost, availability of network resources to transport large amounts of data with stringent quality of service requirements, and coexistence with other existing services using the same resources. This project's objective is to develop a cognitive and reflective network of mobile sensors capable of minimizing their impact on critical city communication resources using a notion of intelligence permeating the layered communication, processing, and sensing infrastructure that characterizes urban environments. The successful realization of this project will contribute significantly to fulfilling the promise of real-time Urban Internet of Things (IoT) systems in the context of limitations imposed by technology and cost. The project also includes a multi-tiered education, mentoring, and outreach plan to train the next generation of IoT systems designer and professionals.The proposed system enhances the ability of individual sensors to make navigation decisions with an intelligent layered architecture capable of providing real-time feedback on the usefulness of their produced data from the perspective of a global computational objective. The real-time adaptation process, then, is expressed over the layers of the architecture, where the agents dynamically learn utility models and control at different geographical and temporal scales. These utility models are used to orchestrate the mobile sensor's navigation within the city to enable detection and monitoring of events and dynamic processes. The outcome of this project is the first architecture of this kind where cognition and intelligence spread across scales of an urban sensing, communications, and processing infrastructure. Furthermore, the system envisioned in this project represents one of the few and most innovative examples of edge computing architecture, where the availability of low-delay processing is used to optimize the system's operations. The construction of a framework for such a complex layered scenario, which involves devices with different sensing and computation capabilities, presents inherent technical challenges which will be addressed by producing several innovations in the area of distributed and hierarchical learning and robot navigation