Energy-Efficient Query Optimization for Sensor Networks using time series forecasting

In this paper a novel approach for optimizing the execution of queries over sensor networks based on time series forecasting is introduced. For improving the performance of data collection, a probabilistic approach, which is the linear prediction model is utilized. In such a ystem, some super nodes and normal sensor nodes cooperate with one another for acquiring data and processing queries. Every sensor constructs time series model based on previous data and transmits only the parameters of the model to the super node. Then, the super-node redicts the future value without directly communicating with sensors and answer queries with approximation. Thus, the quantities and sizes of messages transmitted among sensor nodes can be greatly reduced, so prolonging the lifetime of a sensor network. For verification purposes, the proposed method is compared with real sensor data based on experimental results. Through extensive simulations, the effectiveness of the proposed method is validated. As a result, it is demonstrated that the proposed approach reduces energy consumption on queries.

Keywords:

Wireless sensor database system , Query processing , time series forecasting