A theoretical data management approach for sensor networks

The large-scale sensor network is developing rapidly, thereby more and more efficient methods and algorithms for data management in sensor network are proposed. Traditional database technologies have been moved to this new area. The sensor network community also has developed various technologies for the specific network. This paper proposed a new research issue and method which provides a virtual table for each measurement attribute of sensor nodes. All these virtual tables compose of a virtual database.We develop three steps to produce more accurate data while cost less energy consumption. We can predict the value and estimate these data using EKF. Finally, the record will be revised according to the new reading and the cross attribute model.