Improving Responsiveness of Sensor Webs

In this thesis, we present a sensor network programming platform using fuzzy logic. A fuzzy controller model is used to dynamically control the rate of observation of environmental variables. Differing rates arise from changing environmental conditions. Inferencing using linguistic rules provides a compact, human friendly way to represent the knowledge base for controlling environmental sensor networks. Our platform is illustrated with an example having two rules and three environmental variables.;Our implementation and simulated experimental results show that it is feasible to apply this programming model to wireless sensor networks. The power consumption overhead of fuzzy Sensor Web Language (SWL) using --2.7 dBm transmission power was approximately 15% when compared with Crossbow's benchmark using 0 dBm transmission power without fuzzy SWL. The RAM and ROM usage of fuzzy SWL increased linearly as the number of rules increased. A novel aspect of this research is the addition of time factors in the fuzzy rules. This permits the fuzzy control system to better model and adapt to slowly changing environmental conditions.