Un índice espacio temporal para puntos móviles basado en estructuras de datos compactas

Abstract

Spatio-temporal databases were designed to manage very large datasets of spatial objects, whose location and/or shape evolves with time, and whose changes are relevant to the domain of application. Air traffic control systems, fleet management, migratory birds and other animals are just some examples of this type of systems. Much research on spatio-temporal databases has been devoted to spatial access methods and efficient indexes for secondary memory. However, just a few works tackle the efficient access in main memory. In the field of Information Retrieval, some strategies have arisen to develop new data structures and efficient algorithms in terms of space usage that also keep efficient access times. These have been called Compact Data Structures. These data structures are very space-efficient (getting good compression ratios), while supporting access queries to the data efficiently without decompressing them. Similarly, several data structures that use compression techniques have been developed in the context of spatial access methods. However, to the best of our knowledge, there do not exist previous works on the use of compact data structures for spatio-temporal databases. Therefore, this thesis proposes the use of compact data structures in the context of spatio-temporal databases and, in particular, to index moving objects (represented by spatial coordinates). As a result, a compact self-index is proposed to solve timeslice, time-interval, trajectory and 𝑘-nearest neighbor queries. An experimental evaluation shows that our proposal is efficient to support such queries, while using few space.