The Study On Data Aggregation For Wireless Sensor Networks

The development of micro-electronics,computation and wireless communication has promoted the production and development of wireless sensor network.In many applications users want to get the data of the whole network to the best of its abilities,but the existing data aggregation methods can only get the single eigenvalue by queries to save energy. So it is essential to design a new method to get the whole data with regional information to the best of its abilities while ensuring the life-span of the network.In this paper we introduce the structure,the characteristic of sensor network and the existing typical data aggregation method.We propose a new data aggregation method correlating with the dimensional events.This new method combines with the LEACH protocol to set up two dimensionalities q-digest.The new method first divides the network to many areas which has similar data distribution.Then signs these areas orderly to add the regional informations to the bottom clusters.The parent clusters compress the q-digest with regional information until the sink node gets the whole data of the network and corresponding regional information to complete the assignment of data distribution of the netwok with approximate location information.In this work we design the q-digest for the one level cluster and the multilevel clusters respectively, then analyse the error bound and the transmission.Also we propose the concept of the confidence factor-θ_f andθ_z which can scale the error of q-digest accurately.We simulate our two dimensionalities q-digest algorithm on the precision,transmission and residual power.The result shows that the transmission of q-digest are one fifth of the list.In the list structure.after two collection will have a disabled node because of the nodes near the sink have a heavy load,but the q-digest will have a disabled node after at least 100 queries. It can get the whole data of the network with corresponding regional information to the best of its abilities.Besides it is more suit for large network,the bigger the network is,the better the q-digest is.