Flexible Location - based Spatial Queries

A model for representing and evaluating flexible Location-Based Spatial Queries (LBSQ) is proposed. In a LBSQ the selection condition is generally a constraint on the distance of the objects in the database (instances) from the user location. Such queries are becoming more and more useful in location-based services such as those provided by cell-phones, Wireless LAN and GPS technologies. However their usefulness is limited by the inability of current systems to represent and manage the imprecision often characterizing the knowledge of the user’s and instances’ locations. In this contribution we propose a fuzzy model of flexible LBSQs in which either the user location, or the instances locations or the selection condition itself or even all of them are imprecise. To define a unifying approach in all cases of imprecision we generalize the notion of the Minkowski sum within fuzzy sets and apply it to combine the (imprecise) user location with the (soft) query condition. This way we derive the actual soft constraint with respect to the user’s location. The instances relevant to the query are those whose locations are included in the actual soft constraint representation to some extent.