Self-Adaptive Selective Sensor Network Querying

The reduction of power consumption during the deployment and operation of sensor networks has commonly been recognized as a key challenge. Many proposals have been put forth to save power by taking advantage of the inherent redundancies in sensor network´s operation by minimizing the number of agents active in answering a query at any point in time. The highest level of power saving can be obtained when approximate query results are acceptable and the selection of active agents takes into consideration the intensity and speed of the event being monitored. A larger number of agents can cooperate during high intensity periods to ensure accuracy, while a lower number of agents is sufficient during quiet periods. In this paper, we propose an approach for self-adaptive selective querying. We introduce a set of metrics that allow each data sink to gauge the level of activity in its environment and adjust its querying strategy and intensity accordingly. We show experimental results and discuss future plans.