Reasoning Under Situation Awareness in a Temporal Database

Situation awareness refers to the capability of systems to perceive an existing or predicted context that determines the values of variables in a changing environment. Despite the enhanced support for managing temporal data, current database systems still lack mechanisms for handling highly dynamic situations in which data may change frequently. We present first results from an ongoing research project investigating these missing database features. In particular, we identify (i) the requirements for representing complex spatio-temporal data, (ii) the reasoning capabilities needed for detecting valid relationships between situations, and (iii) the operators necessary for supporting situation-based reasoning. Our investigations are based on a perception concept, which comprises interval timestamped data derived from observed events and processed using the sequences semantics. Perceptions provide a high level (and qualitative) description of past and current situations, complemented by projections into the future.