Spatio-temporal databases are becoming increasingly important due to the emergence of numerous applications that demand efficient management of moving objects. Supporting objects with dynamic behavior requires significant extensions from traditional spatial and temporal databases. This thesis discusses the indexing and querying issues in optimizing spatio-temporal processing. Specifically. after surveying the existing access methods, we derive efficient cost models that accurately predict their performance (in terms of structure sizes, query costs, etc). Motivated by their deficiencies revealed from the models, an alternative index, the MV3R-tree, is proposed to combine the advantages of previous structures. Furthermore, we formulate the concept of time parameterized queries (TP), which...[ Read more ]

Spatio-temporal databases are becoming increasingly important due to the emergence of numerous applications that demand efficient management of moving objects. Supporting objects with dynamic behavior requires significant extensions from traditional spatial and temporal databases. This thesis discusses the indexing and querying issues in optimizing spatio-temporal processing. Specifically. after surveying the existing access methods, we derive efficient cost models that accurately predict their performance (in terms of structure sizes, query costs, etc). Motivated by their deficiencies revealed from the models, an alternative index, the MV3R-tree, is proposed to combine the advantages of previous structures. Furthermore, we formulate the concept of time parameterized queries (TP), which can be integrated with most spatial query types to obtain predictive information based on the current knowledge about objects’ movements. In particular, TP queries can be processed as nearest neighbor search by defining appropriate distance functions. Based on this we extend the discussion to the continuous query that is also very important in spatio-temporal applications. Finally we discuss aggregate processing in spatial and spatio-temporal databases, proposing several multi-tree indexes.