Sensor nodes and Wireless Sensor Networks (WSN) enable observation of the physical world in unprecedented levels of granularity. A growing number of environmental monitoring applications are being designed to leverage data collection features of WSN, increasing the need for efficient data management techniques and for comparative analysis of various data management techniques. My research leverages aspects of fuzzy database, specifically fuzzy data representation and fuzzy or flexible queries to improve upon the efficiency of existing data management techniques by exploiting the inherent uncertainty of the data collected by WSN. Herein I present my research contributions. I provide classification of WSN middleware to illustrate varying approaches to data management for WSN and identify a need to better handle the uncertainty inherent in data collected from physical environments and to take advantage of the imprecision of the data to increase the efficiency of WSN by requiring less information be transmitted to adequately answer queries posed by WSN monitoring applications.In this dissertation, I present a novel approach to querying WSN, in which semantic knowledge about sensor attributes is represented as fuzzy terms. I present an enhanced simulation environment that supports more flexible and realistic analysis by using cellular automata models to separately model the deployed WSN and the underlying physical environment. Simulation experiments are used to evaluate my fuzzy query approach for environmental monitoring applications. My analysis shows that using fuzzy queries improves upon other data management techniques by reducing the amount of data that needs to be collected to accurately satisfy application requests. This reduction in data transmission results in increased battery life within sensors, an important measure of cost and performance for WSN applications.
