Trustworthiness assessment in mapping urban accessibility via sensing and crowdsourcing

In this work we present the trustworthiness assessment in mPASS (mobile Pervasive Accessibility Social Sensing), a location and context aware system that collects data from crowdsourcing and sensing in order to map urban and architectural accessibility. The fusion of heterogeneous urban sources allows mPASS to provide users with personalized paths, computed on the basis of their preferences and needs. To perform this task, the system needs a set of georeferenced data that is dense enough and trustworthy enough to avoid false positives and negatives, e.g. to prevent users from encounter on their path an unknown barrier or a non-existing facility. In order to reach this goal, we propose a trustworthiness assessment which combines accuracy of sensors, source credibility of the crowd and the authority of experts. To evaluate the effectiveness of our trustworthiness assessment, we conducted a set of simulations by exploiting OSM data and we have obtained interesting results.