On continuous monitoring top-k unsafe moving objects

With the wide usage of location tracking systems, continuously tracking relationships among moving objects over their location changes is possible and important to many real applications. This paper proposes a novel continuous location-based query, called the continuous top-k unsafe moving objects query or CTUO. This query continuously monitors the k most unsafe moving objects, where the unsafety of an object (protectee) is defined by the difference between its safety requirement and the protection provided by protection forces (protectors) around it. Compared with the traditional top-k queries where the score of an object represents its own characteristics, CTUO describes the relationships between protectees and protectors, which introduces computational challenges since naively all objects should be inspected to answer such a query. To avoid this, two efficient algorithms, GridPrune and GridPrune-Pro, are proposed based on the basic pruning technology from the Threshold Algorithm. Experiments show that the proposed algorithms outperform the naive solution with nearly two orders of magnitude on I/O savings.