A context-sensitive infrastructure for coordinating agents in ubiquitous environments

The combination of contextual information about the real world (e.g., collected by sensors) with information coming from the virtual world (e.g., the Web 2.0), may represent an enormous enrichment particularly for services organized and provided by agents in ubiquitous environments. To address the challenging need for coordination in such environments, and to provide the user a high-level of service quality, an engineered approach to exploit such information is required. Such an approach should generate added-value by offering means for combining diverse data sources, should allow delivering context-sensitive information and, hence, should promote context-dependent coordination of entities in ubiquitous environments. In this paper, we report the scenario-based analysis of key requirements for ubiquitous environments that we have used as the basis for the design of our proposal. Following, our proposal for the Ubiquitous Coordination Model (UbiCoMo) and its associated infrastructure is detailed. The UbiCoMo model covers an expressive data model based on four-field tuples to represent contextual information, data distribution and management concepts based on tuple spaces, and the integration of coordination patterns to resolve reoccurring coordination problems in ubiquitous scenarios. The UbiCoMo infrastructure integrates the fundamental mechanisms for agent-based coordination in ubiquitous environments, and is well-suited to provide the specific means required to offer high-level services and context-sensitive functionalities. A concrete ubiquitous application, the living diary, is assumed as a case study both to illustrate the requirements analysis and to exemplify the usage and the suitability of UbiCoMo.