By Gon?alves, N.; Jos\'e, R.; Baquero, C.
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
The information infrastructure that pervades urban environments represents a major opportunity for collecting information about Human mobility that would be very important across many application domains. However, this huge potential has been undermined by the overwhelming privacy risks that are associated with such forms of large scale sensing. In this research, we are concerned with the problem of how to enable a set of autonomous sensing nodes, e.g. a Bluetooth scanner or a Wi-Fi hotspot, to collaborate in the observation of movement patterns of individuals without compromising their privacy. We describe a novel technique that generates Precedence Filters and allows probabilistic estimations of sequences of visits to monitored locations and we demonstrate how this technique can combine plausible deniability by an individual with valuable information about aggregate movement patterns. The results provide a promising step towards the application of new stochastic techniques in large scale sensing.