Querying Imprecise Data in Sensor Databases

Sensors are used to monitor some physical phenomena such as contamination, climate, building, and so on. The sensors collect and communicate their readings to the sensor databases for making decisions and answering various user queries. Due to continuous changes and possible errors in these values, the data values recorded in sensor databases may differ from the actual status. Queries using these values can yield incorrect and misleading answers. In order to manage the imprecision between the actual sensor value and the database value, the framework representing imprecision of sensor data has been proposed, in which each data value is represented as an interval. In this paper, we examine the situation when answer imprecision can be represented qualitatively and quantitatively. In particular, we propose two new kinds of imprecise queries called Top-k imprecise query and imprecise threshold query. Also, we investigate techniques for evaluating the qualitative and quantitative queries.