Este TFM se ha desarrollado dentro del periodo de investigación tutelada del Máster en Soft Computing y Sistemas Inteligentes, en el año 2012.
In this work we study systems that allow detection and recognition of human activities. We present a model to represent human behaviour in Smart Spaces with semantics; more concretely, with ontologies. The domain is focused within a working environment, such as the office domain. This model will be used to learn, recognize and monitor behaviours, and will be simple enough for those behaviours to be easily programmed not only by a programmer, but also by a non expert. In order to achieve this, the model is graphical, with the aim of equipping it with a broader usability in the end-user application. Context-aware reasoning allows to infer novel information from atomic data. The proposal also contains mechanisms to face the uncertainty that the environment around the user may encounter. Our research topic falls under the context of Ambient Intelligence (AmI), which plays an essential role on integration and treatment of information from sensors that describe the user activity in the environment for a later transfer and processing, in real-time, by Artificial Intelligence (AI) models. The final aim consists of tracking humans and events or behaviours, such as their location or activity, to generate reminders or alarms reacting to, for instance, forgotten actions or potentially hazardous situations.