Ambient assisted living ecosystems comprising both “intelligent” living spaces based on distributed sensor/actuator networks and assistive robots that provide safe, pervasive, and long-term interactions with people at their household need to be programmed with a basic set of capabilities that can grow and adapt over long periods of time (e.g. for many weeks, several years), by learning and adapting to individual needs, habits and preferences of the users. The research team is pursuing research to make those connected and smart environments adapt to the users, as regards to different aspects of their lives: comfort, daily activity management, safety, security, socialization. The research also focus in localized and distributed context, for example, to what extent does the information from a person’s typical routine should affect the environment behavior to maximize a specific goal.