CSR-EHS: Novel Mobile and Distributed Embedded Systems for Pervasive Computing Applications

普适计算系统即将给我们的世界带来革命性的变化--计算将体现在日常物品中，而不是“传统”的计算机中，并将无缝地改善我们的生活。这种嵌入和成本考虑表明，普适计算应用程序将由一组动态的系统组件(智能对象)组成，这些组件在分布式交换矩阵上执行，其中包括共享存储、计算和传感资源、适应环境变化以及机会主义地合作和利用第三方服务。与此同时，RFID技术继续其快速发展，以实现标签、存储器(即，读/写存储器)的成本效益嵌入，并因此能够在物体的环境和空间内嵌入上下文。因此，这些技术为探索普适计算特有的系统级挑战提供了一个具体的环境。这项研究为下一代普适计算应用的新的系统级技术和设计方法奠定了基础。它包括三个关键的冲刺。首先，讨论了基于RFID的新型分布式存储子系统的组织和管理。关键思想是，这些可以用作公共资产，由多个应用程序共享，利用一组匿名移动智能对象在情景中收集和高效传播上下文信息。第二个重点是维护这些基于RFID的分布式存储子系统之间的数据完整性和一致性。最后一个重点是研究如何为新的普适计算应用生成和存储上下文，包括对生成和捕获分布式时空上下文数据的不同策略的特征和可伸缩性的正式研究。

Pervasive computing systems are on the brink of revolutionizing our world - computation will be embodied in everyday objects, not 'traditional' computers, and will seamlessly enhance our lives. This embedding, and cost considerations, suggest that pervasive computing applications will consist of a dynamic set of system components (smart objects) executing on a distributed fabric consisting of shared storage, computing, and sensing resources, adapting to environmental changes, and opportunistically cooperating and exploiting third party services. At the same time, RFID technologies continue their fast paced evolution towards enabling cost effective embedding of tags, storage (i.e., read/write memories) and thus context within the environment and 'space' of objects. As such, these technologies provide a concrete setting in which to explore the system-level challenges unique to pervasive computing. This research lays the foundations for novel system-level technologies and design methodologies for next generation pervasive computing applications. It includes three key thrusts. First, it addresses the organization and management of novel RFID-based distributed memory subsystems. The key idea is that these might be used as a public asset, shared by multiple applications, leveraging a collective of anonymous mobile smart objects to gather and efficiently disseminate in situcontextual information. The second thrust focuses on maintaining data integrity and consistency across such RFID-based distributed memory subsystems. The last thrust investigates how to generate and store context for novel pervasive computing applications, involving a formal study of the characteristics and scaling properties of different policies to generate and capture distributed space-time contextual data.