An AmI-Enabled OSGi Platform Based on Socio-Semantic Technologies
Ana Fernandez, Rebeca P. Dı́az Redondo, Jose Juan Pazos-Arias, Manuel Ramos, Alberto Gil, Jorge Garcia · InTech eBooks · 2010
The primary reference architecture in the OSGi specification is based on a model where a operator manages a potentially large network of service platforms. It assumes that the service platforms are fully controlled by the operator and used to run services from many different services providers. In this scenario, we have shown that the actual service discovery mechanisms in OSGi is insufficient. In the pursuit of a really open and interoperable residential gateway, we propose the semantic description and discovery of the services in the OSGi domain. At this respect, we have defined OWL-OS, a sub-ontology of OWL-S which allows making a simple semantic search of services based on a categorized structure. Despite we propose operations-at-home as the primary structure to classify the OSGi services, OWL-OS allows an OSGi service to be semantically described according to different ontological structures. These ontological structures would be downloaded on demand from the service provider. Finally, note the Semantic OSGi Framework enhance the OSGi standard, without breaking it; i.e. any non-semantic bundle can work properly within this framework, although it is not able to take advantage of the semantic reasoning for service obtaining. Moreover a clear benefit of the new Semantic OSGi platform is the possibility of supporting the automation of OSGi services composition (Díaz Redondo et al., 2007). In the field of service selection, it would be possible to supplement the Semantic Registry with different specialized software agents which takes into account other factors (ambient intelligence) to automate the service selection. In this paper we have investigated the potential of combining a personalization agent and a context management system. On its own, the personalization agent contribute to decisions on service selection by using a preference model storing historic information about preferred devices and services. On the other hand, the context management system allows the OWL-OS Framework to be aware of the particular context (time, location, mood, etc) and to react by discovering services which meet that context. Beyond their independent behavior, connecting the personalization and context-aware agent across an ontological base, OWL-OS in our case, allow the OWL-OS framework to support scenarios as the one included in this paper. In general, scenarios where the selection of the service depends not only on who uses the service but when, where or even why uses the service. Although this paper only discusses two forms of context-aware personalization in the smart home (service selection and parametrisation), some other forms are possible with the general preference model we propose. For instance, provider selection can be personalized by capturing the preferences of the user. Apart from the general approach, one of the main contributions of this work is the effort to provide a flexible context model which fit in well with an extremely heterogeneous environment. With this premise, we provide a collaborative approach to context, in fact we can say that the solution is collaborative in two senses. Firstly, it is a collaborative model of context since all the devices, services, sensors, inhabitants, etc. at home form a pluriarchy without any controlled vocabulary. Secondly, it is also a collaborative model for preferences since the service selection takes into account not only the preferences of the user but even the preferences of a smart home community, possible a community established by the service provider. At this respect, in this paper we have explored algorithms of service selection which rely on tag cloud