Ambient assisted living is a key enabler to increase the quality of life, and its importance is going to increase in the next decade due to a significant increment in the number and in the percentage of the elderly population. Many solutions have been proposed to compensate the age related physical limitations and to guarantee autonomy and safety to the elderly people, allowing them to continue their life in their own environment. There are several open issues to solve in order to enable the practical implementation of these solutions. One of them regards the intrinsic difficulties to create a network between a personal smartphone and the correct physiological sensors, in the presence of multiple sensor devices, where the automatic choice of the correct sensor is not obvious. In this paper we propose a solution to this problem, based on an automatic pairing that relies on the instantaneous variability of the heart rate of each subject. We evaluate the technique by collecting real data and showing some promising performance accuracy results, opening new research opportunities in this area.