A Hybrid Framework to Program Smart Environments

Smart environments improve the user experience by customizing the environment according to her/his preferences. Preferences are either captured by agent learning or specified through user-centric programming. Autonomous agents use artificial intelligence to implicitly learn the reactions to different stimulus without user control over what or when to learn. Conversely, user-centric techniques enable users to explicitly associate the appropriate actions with specific conditions in a teaching session. This project presents a hybrid framework that utilizes behavior-based AI to customize smart environment settings while giving the user control over the high level environment behavior. The framework uses a voice recognition interface supported by an underlying programming engine to improve usability, flexibility and correctness.