Handling location uncertainty in probabilistic location-dependent queries

Location-based services have motivated intensive research in the field of mobile computing, and particularly on location-dependent queries. Existing approaches usually assume that the location data are expressed at a fine geographic precision (physical coordinates such as GPS). However, many positioning mechanisms are subject to an inherent imprecision (e.g., the cell-id mechanism used in cellular networks can only determine the cell where a certain moving object is located). Moreover, even a GPS location can be subject to an error or be obfuscated for privacy reasons. Thus, moving objects can be considered to be associated not to an exact location, but to an uncertainty area where they can be located. In this paper, we analyze the problem introduced by the imprecision of the location data available in the data sources by modeling them using uncertainty areas. To do so, we propose to use a higher-level representation of locations which includes uncertainty, formalizing the concept of uncertainty location granule. This allows us to consider probabilistic location-dependent queries, among which we will focus on probabilistic inside (range) constraints. The adopted model allows us to develop a systematic and efficient approach for processing this kind of queries. An experimental evaluation shows that these probabilistic queries can be supported efficiently. (C) 2017 Elsevier Inc. All rights reserved.