A model for mobile-instrumented interaction and object orchestration in smart environments

Resumen

The proliferation of the smartphones has given a considerable boost to the spread of the smart objects and the consequent creation of smart spaces. Smart objects are electronic devices that are able to work interactively and autonomously, usually preserving the interaction metaphor of their non-electronic counterpart. Through a network interface, they can cooperate with other objects: their strengths do not lie in their hardware, but in the capabilities to manage interactions among them and in the resulting orchestrated behaviour. Smart spaces are environments composed of smart devices, where they work together, producing some behaviour of benefit to the dwellers. The current workflow requires that the user downloads an application from a digital distribution platform for each smart object he wants to use. This model converts the smartphone in a control centre, but it limits the potentialities of the objects. Devices are connected in a one-to-one network with the smartphone, a configuration that prevents the direct communication among the objects and that puts the responsibility to coordinate the objects in the hands of the smartphone. Moreover, there are only a few frameworks that permit the integration of several applications and the creation of complex behaviours that involve many objects from different manufacturers. The first challenge considered in this thesis is to propose a new workflow that permits to integrate any kind of smart device in any behaviour of the smart space. The workflow will include the discovery of the new objects and their configuration, without the need of downloading a new standalone application for every object. It will provide to the user a simple configuration tool to create personalized behaviours (scenes), based on the event-condition-action paradigm. Finally, it will automatically orchestrate the smart devices to produce the desired behaviours of the environment. Smart spaces are thought to behave in a personalized way, adapting to the particularities of their inhabitants. “Personalization” is about understanding the needs of each individual and helping satisfy a goal that efficiently and knowledgeably addresses each individual’s need in a given context. Thus, the second challenge tackled in this thesis is to move forward on how to evolve smart spaces from customizable to personalized environments. The third open issue considered in this research is how to make portable the personalized configuration, i.e. how to enable a user to commute among different smart spaces preserving and adapting his personalized settings to each of them. Both personalization and portability will be included in the tool to automatically aid the user to gain a full and transparent control over the environment. Solutions to tackle device fragmentation, interoperability, seamless discovery, scene modelling, orchestration and reasoning are needed to achieve these goals. In this context, the contributions of the thesis can be summarized as follows: - The definition of a workflow that permits to personalize and control a smart space using the event-condition-action paradigm. - The design of a model to describe any kind of smart object and its capabilities. - The extension of the previous model to describe a smart space as a composition of its smart objects. - The application of the model to the workflow, reinterpreting the main Object- Oriented Programming features and using them to describe the interactions between objects and the recommendation process. - The proposal of an architecture that implements each step of the workflow and its relations with the model. - The proposal, development and evaluation of service concepts in real smart space settings.