High Reliable In-Network Data Verification in Wireless Sensor Networks

Abstract

In order to provide efficient and suitable services for users in a ubiquitous computing environment, many kinds of context information technologies have been researched. Wireless sensor networks are among the most popular technologies providing such information. Therefore, it is very important to guarantee the reliability of sensor data gathered from wireless sensor networks. However, there are several factors associated with faulty sensor readings which make sensor readings unreliable. In this paper, we classify faulty sensor readings into sensor faults and measurement errors, then propose a novel in-network data verification algorithm which includes adaptive fault checking, measurement error elimination and data refinement. The proposed algorithm eliminates faulty readings as well as refines normal sensor readings, to increase reliability. Also, to achieve scalability of sensor networks and minimize network overhead, the proposed scheme involves a distributed implementation in a local area. The simulation study shows that the in-network data verification algorithm is highly reliable and its network overhead is very low compared to previous works. Reliability and overhead is improved by a maximum of 10---30% and 70%, respectively.