Leveraging fuzzy query processing to support applications in wireless sensor networks

In this paper, we describe a fuzzy query processing approach to support application development in sensor networks. Using a fuzzy query, an application programmer can provide a linguistic and semantic specification of the desired data, eliminating the need to specify explicit and exact thresholds as part of a query. The returned fuzzy query results are each associated with a degree of membership measurement that indicates how closely each returned data value matches the semantic intent of the fuzzy query, providing applications with additional information that can be used to reason about the query result. Our approach to in-network fuzzy query processing allows for each sensor node to tailor its evaluation of a fuzzy query; this feature allows for consideration of micro-environments embedded within the sensor network that can impact how individual sensor data values should be interpreted with respect to the semantic intent of the query. To demonstrate that a fuzzy query processing approach is feasible, we use an application scenario to evaluate the implementation of our fuzzy query processing system in a simulated sensor network environment; results show that precision and overhead for our approach are comparable to traditional query processing.