Master 2 subject: Control of Smart Environments and Probabilistic Models

Smart environments, like smart homes or smart buildings, are environments designed to react and adapt themselves to their environmental context : users behavior, time or climate change, etc. The prospects of such environments goes from improved safety and wellbeing to ecological concerns. These environments are supplied with sensors and actuators, and with computing facilities allowing the design of autonomic loops, implementing smart policies. The increasing size and complexity of such systems, together with the need of safety properties for the user (e.g., physical safety or privacy), lead to the need of design methodologies able to answer to these safety concerns. Instead of the classical implementation/validation process, we propose to use a method stemming from control theory : discrete controller synthesis [1]. However, this method does not allow easily to express and use probabilistic events or degraded mode.