An Architecture for Interactive Query Refinement in Sensor-based Information Fusion Systems
Giovanni Casella, Gennaro Costagliola, Erland Jungert, Li Xin, Tobias Horney · 2004
An architecture for query execution and interactive query refinement on heterogeneous data from spatially distributed sensors is proposed. The objective is to define a system in which it is always possible to add new data sources and new algorithms for data processing and user query execution. The query language that forms the basis for query formulation is called ΣQL. Ambiguous and vague query can be interactively refined as well. An ontological knowledge base to model fundamental architecture components and their relations is introduced. The method to choose algorithms that can execute user queries based on available data, the Pequliar interactive system, the Reasoner and the user interface are described. The Reasoner automatically or semi automatically carries out an iterative information fusion step by means of rules that are either automatically launched by the Reasoner or defined by the user. If a rule is defined by the user the Reasoner can learn the new rule and use it in later queries. Fusion query examples are also presented. User interface enables the user to compose compound queries from elementary queries using query operators, presents the query results and gives the user a qualitative evaluation of the results. 1.