Query processing on historical uncertain spatiotemporal data
(Y. Tao)

Spatiotemporal databases (STDB) are the core of numerous applications (e.g., transportation control) that require continuously tracking the locations of a large number of moving objects (e.g., vehicles). In practice, an object is equipped with a GPS device, and reports its location to a server periodically through a wireless network. Spatiotemporal data is inherently uncertain, because a server does not have the precise location of an object at an un-reported timestamp. Although several models have been proposed to capture uncertainty, query processing on historical uncertain spatiotemporal data has not been studied. The first objective of this project is to explore novel query types that facilitate probabilistic information retrieval in such environments. As a second step, we aim at developing a general access method, which is applicable to all the uncertainty models, and can efficiently answer various types of probabilistic queries. Our research outcome has significant importance for analyzing and improving the traffic conditions in Hong Kong. Furthermore, the techniques developed in this project can also enhance general applications that require effective retrieval of uncertain locations of moving objects in the past.