Ambient Assisted Living (AAL) pursues providing an autonomous and satisfactory life to people through technology, independently of their actual conditions. Its developments usually require testing protoypes with real users in Living Labs (LL). This makes projects expensive. Virtual LLs (VLLs) try to address theses issues by using simulations for requirements elicitation and the initial testing of solutions. These simulations frequently require considering social aspects, e.g. relationships,

more »

... ure, or decision making. These are recurrent and quite applicationindependent aspects for AAL. Our work proposes social properties as patterns that represent these aspects and that can be plugged-in in simulations. The knowledge for these properties is extracted following the Activity Theory (AT) paradigm from Social Sciences. Their specification uses models and transformations (e.g. to generate other models or code) following Model-Driven Engineering (MDE) practices. This faciliates their understading and use in simulation development. A case study on AAL for ageing illustrates the approach.