The State Data Aggregation Research In Wireless Sensor Networks

Wireless sensor network is a distributed and self-organizing network system, which is form of a large number of ubiquitous tiny sensor nodes with wireless communications and computing capability, this is a "smart" system which can complete the assigned mission according to the environment. Because the energy of nodes is supplied by battery and their working environment is complex, this makes it difficulty to supply energy to nodes, so in order to maintain network's longer working time, the limited energy of nodes becomes a great bottleneck of wireless sensor network. While because the energy consumption of the nodes for transmitting data is much larger than the data processing, so we can increase the data processing capacity, reduce the volume of data transmission to reduce power consumption, then we need to carry out data aggregation. In this paper, two kinds of data aggregation methods have been proposed, which are about the different state data's characteristics of nodes.⑴The energy reporting aggregation mechanism based on forecast and difference- threshold.Even though the working environment of nodes is an uncertainty and volatility, their energy consumption has some law, because working state is relatively stable in a period of time, which provides some preconditions for establishing energy model. This method establishes energy forecast model, and then reports state data by difference-threshold method, which can decrease the counts of sending data. The experiments shows that this method can significantly reduce the count of data transmission, decrease energy consumption, extend the lifespan of the network.⑵The n-preindexing coding compression algorithm based on clustering.Because the working environment is similar in a cluster, so the residual energy and link quality are also similar or identical, which make a large number of redundant data in the cluster. In order to eliminate the redundancy, this paper presents the n-preindexing coding compression algorithm based on clustering, which also designs optimizing packet format, avoiding reporting common node's ID in a cluster. The analysis about the experimental datas shows that this method can reduce the state data of nodes about 30%.