CAREER: Towards Spatial Data Systems Support for the Internet of Things

物联网（IoT）代表一个设备网络，每个设备都配备了各种传感器，收集有关环境的数据。自2012年以来，物联网一直在快速增长，约有90亿台设备，其中2018年约有200亿台设备，到2020年将有超过300亿台设备连接到物联网。物联网设备可以安装在建筑物或交通路口等静态位置，以收集有关城市空气污染，水质或交通的数据。智能手表等设备也可以连接到移动物体（人）上，以监测人的运动模式、心率或燃烧的卡路里。在这两种情况下，由IoT设备生成的数据具有表示感测到的观测的地理空间位置和时间的空间和时间属性。此外，各种设备测量不同的观察结果（例如，温度、声音、速度等）。该项目的目标是构建创新的可扩展技术，无缝连接从各种地理分布的物联网设备收集的数据，从而有效地管理和分析各种空间和时间尺度上不断增长的物联网数据。该项目的成果将为政策制定者、科学家、企业和公民提供一种工具，以便在交通、环境经济、公共安全和健康等各种应用中更好地利用物联网数据并从中提取价值。为了整合研究和教育并实现社会影响，项目团队还将启动“城市数据科学倡议”，并在亚利桑那州的滕佩市进行试点。该项目旨在通过分析从物联网设备收集的数据，让当地社区和当地学生参与解决许多城市问题的工作。该项目的总体目标是开发图形数据库系统技术，该技术可以有效地存储，管理和执行链接物联网数据的实时空间/时空图形查询，并支持大规模物联网数据的可扩展处理。为了实现这一目标，该项目的研究工作包括设计和开发新的空间数据存储、索引、处理和管理技术，以适应物联网设备收集的数据量和速度的不断增加。与现有的大空间数据系统和数字框架相反，一种新的物联网数据抽象方法将扩展最近开发的大空间数据系统，为物联网应用程序的开发提供应用程序编程接口（API）。新开发的方法不仅考虑到数据的空间分布，而且还考虑到每个传感器产生的信号的物理和数学特性。此外，新的物联网抽象方法包括新的查询运算符和查询优化策略，可以有效地评估混合工作负载，其中涉及传统的空间和时空数据处理以及对物联网数据的数字信号处理操作。新方法还通过设计一个中间件来弥合物联网设备和空间数据系统之间的差距，该中间件将物联网设备的数据流集成到中央系统，并满足访问此类物联网数据的应用程序的要求。该项目的另一个成果是一个图形查询处理器，它可以优化和评估通用的空间谓词，如空间范围，连接和K近邻（KNN）以及在真实的时间内对链接的物联网数据发出的图查询中的时间谓词，即使新事物和观察定期添加到物联网图中。该奖项反映了NSF的法定使命，并被认为值得通过以下方式提供支持：使用基金会的知识价值和更广泛的影响审查标准进行评估。

The Internet of Things (IoT) represents a network of devices, each equipped with a variety of sensors that collect data about the environment. The IoT has been growing rapidly since 2012 with about 9 billion devices, including around 20 billion devices in 2018 and reaching more than 30 billion devices connected to the IoT by 2020. An IoT device can be installed in a static location such as a building or a traffic intersection to collect data about air pollution, water quality, or traffic in a city. A device, such as a smart watch, can also be attached to a moving object (human) to monitor a person's movement pattern, heart rate, or calories burned. In both cases, data generated by an IoT device possess spatial and temporal attributes that represent the geospatial location and time of the sensed observation. Furthermore, various devices measure different observations (e.g., temperature, sound, speed, etc.). The aim of this project is to build innovative scalable technologies that seamlessly connect data collected from various geographically distributed IoT devices that can effectively manage and analyze the ever growing IoT data at various spatial and temporal scales. The results of the project will provide a tool for policy makers, scientists, businesses, and citizens to better utilize and extract value from IoT data in a variety of applications, such as transportation, environmental economics, public safety and health. To integrate research and education and achieve societal impact, the project team will also kick off the "Data Science for Cities Initiative" with a pilot chapter in the city of Tempe, Arizona. The initiative aims to engage local communities and local students in working on solutions to many urban problems by analyzing data collected from IoT devices.The overarching goal of the project is to is to develop graph database systems technology that can efficiently store, manage, and execute real-time spatial / spatio-temporal graph queries on linked IoT data and support scalable processing of large-scale IoT data. To achieve that, the research effort in this project includes the design and development of novel spatial data storage, indexing, processing and management techniques that scale to the ever-increasing volume and fast rate of data collected from IoT devices. As opposed to existing big spatial data systems and numerical frameworks, a novel IoT data abstraction method will extend recently developed big spatial data systems to provide an Application Programming Interface (API) for development of IoT applications. The newly developed method takes into account not only the spatial distribution of the data, but also the physical and mathematical characteristics of signals generated by each sensor. Furthermore, the new IoT abstraction method includes novel query operators as well as query optimization strategies, which can efficiently evaluate a hybrid workload that involves classic spatial and spatio-temporal data processing and digital signal processing operations on IoT data. The new method also bridges the gap between IoT devices and the spatial data system by designing a middleware that integrates the IoT devices streaming data to the central system with the requirements of applications accessing such IoT data. Another outcome of the project is a graph query processor that can optimize and evaluate general-purpose spatial predicates such as spatial range, join and K-Nearest Neighbors (KNN) as well as temporal predicates in a graph query issued on linked IoT data in real time even with new things and observations being regularly added to the IoT graph.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.