Open infrastructure for data sharing and interoperability between context-aware applications

最近移动和无处不在的计算技术（如全球定位系统（GPS）、无线射频识别系统（RFID）和闭路电视（CCTV））的激增，使人们能够探索情境感知服务，这些服务可以监控和适应用户，并在用户进行日常活动时为他们提供有用的服务。上下文感知计算的现有研究主要集中在三个方面：构建新的应用程序；构建新的技术和工具，以感知和推断环境和更高层次的活动；而软件基础设施的建设有利于这类应用的开发。尽管进行了广泛的调查，但这一领域的研究已经产生了1)数量有限的不能相互操作的独立应用程序；2)一套用于感知和推断上下文和高级活动的技术和工具，这些技术和工具很难互换使用或比较评估。由于目前还没有标准，这些研究人员都使用了上下文信息的本地表示。我们认为，出现这些问题主要是因为缺乏跨各种上下文感知应用程序和工具使用的一致的信息体系结构。为了实现互操作性和互换性，需要在不同的应用程序和工具之间以有意义的方式感知和共享上下文数据。此外，通过创建用于比较各种工具的通用测试数据集，可以促进对传感器、活动推断技术和应用的评估。所提出的研究包括设计、开发和评估用于存储和访问上下文信息的灵活开放的基础设施。

The recent proliferation of mobile and ubiquitous computing technologies (such as global positioning systems (GPS), radio frequency identification systems (RFID), and closed circuit television (CCTV)) has enabled the exploration of context-aware services which monitor and adapt to users, and provide them with useful services while they perform daily activities. The majority of existing research in context-aware computing has focused on three main areas: the construction of novel applications; the construction of new techniques and tools for sensing and inferring context and higher level activities; and the construction of software infrastructures to facilitate the development of this class of applications. Despite extensive investigations, research in this space has produced 1) a modest number of stand-alone applications which cannot interoperate with one another and 2) a set of techniques and tools for sensing and inferring context and high-level activities that are difficult to use interchangeably or evaluate comparatively. Investigators in each of these efforts have used homegrown representations of context information because no standards currently exist. We argue that these issues arise primarily because of the lack of a consistent information architecture used across the various context-aware applications and tools. To achieve interoperability and interchangeability requires that context data be sensed and shared in a meaningful way between different applications and tools. Furthermore, evaluation of the sensors, activity inferring techniques, and applications could be facilitated through the creation of common test data sets for comparing various tools. The proposed research involves the design, development, and evaluation of a flexible and open infrastructure for storing and accessing context information.