A Trustworthiness Model for Crowdsourced and Crowdsensed Data

This work presents mPASS (mobile Pervasive Accessibility Social Sensing), a system that exploits data crowdsourcing and crowdsensing to support urban accessibility. The system aims providing users with personalized paths, computed on the basis of user profiles and of the accessibility facilities/barriers present in the location. To perform this task, mPASS needs a set of georeferenced data dense enough and trustworthy enough to avoid false positives and negatives. With these needs in view, mPASS combines data gathered by users and sensors, with information produced by disability organizations and local authorities. In this paper, we propose a method to evaluate trustworthiness of data provided by the system, taking into account characteristics of the different data sources. We conducted a set of simulations on credibility of sources, obtaining positive results.