A Probabilistic Framework for Localizing People with K-anonymity.
Francesco Buccafurri, Gianluca Lax · SEBD · 2009
In many application contexts the possibility of knowing user’ position inside a limited environment can be very important, since may represent a valid support to provide precious services (think of patient’s care in health assistive environments or localization of prisoners inside a penitentiary). However, we cannot assume in general that the utility of having precise information about user’ location is stronger than the right of keeping private the access to some place and, more in general, the exact movements that users do in the environment. In this paper, besides the standard precise localization issue, we consider also the issue of providing approximate non-deterministic answers about user’s position in such a way that privacy is protected yet maintaining the possibility that a user is found with a small number of attempts. The solution enforces the general concept of k-anonymity, which we adapt in our context in a way different from the classical one. An important aspect of our technique is that it is strongly efficient and implementable with very cheap devices. This is a very relevant issue in pervasive environments where wireless devices with limited processing capability and power have to be utilized.