Research On Data Aggregation Technologies Of WSN

Wireless sensor networks are composed of a large number of sensor nodes which cooperate with each other. Wireless sensor networks are data-centric, with limited battery power of nodes, limited hardware resources, the large number of nodes, self-organization and dynamic topology. Therefore, a primary goal in the design of wireless sensor networks is lifetime maximization.The function of data aggregation in wireless sensor networks is to reduce traffic, reduce network congestion,increase the accuracy of information, delete data that is redundant, inefficient and poor credibility, and at the same time aggregate data which is collected from different nodes so that the amount of data transfer in network can be reduced and energy can be saved effectively. Then the lifetime of the network prolonged. According to information entropy that is contained before or after the data aggregation, We can group aggregation into two types: lossless aggregation and lossy aggregation. According to that the data examples is based on or not semantic of application data, We can group aggregation into two types: application dependence data aggregation, application independent data aggregation and the combination of above two kinds of data aggregation. Based on the operational-level of sensor data, data aggregation technology can be divided into data-level aggregation, feature-level aggregation and decision-level aggregation.The factors that affect the total data transfer include characteristics of monitoring data, expression and network topology, as well as specific applications. The current data aggregation research mainly focus on the application layer and the network layer. The data aggregation-oriented interface has been developed in the application layer and the data aggregation technology integration of routing has been developed in the network layer. Outside the existing protocol layer, the data aggregation technology independent of the application has also put forward,creating the data aggregation layer between the link layer and the network layer.In the application layer, the research on data aggregation technology is most based on data aggregation technology of the query mode, and the most representative is TinyDB which is a query processing subsystem of TinyOS developed by the University of Berkeley.It is a high-level abstract, taking sensor networks as a distributed database,programming on data-centric, providing a simple Tiny-SQL query interface for users, and providing extendable framework model. TAG is a data simple aggregation model based on query.Base on it is whether or not data aggregation considered, wireless sensor networks can be grouped into two types:to address-centric routing and data-centric routing. It first choose the best path to sink for each information source in data-centric driven routing algorithms, and the information from each node will aggregate if meeting in the course of being sent to sink.The key to implement data-centric routing protocol is to integrate data routing and data aggregation. In the sensor network,sink collection of data by data aggregation tree, and if each intermediate node aggregate the received data, the information is maximum extent of aggregation in time.Data-centric driven routing algorithms choose the best path to sink first for each information source node, and aggregate the information that meet in the intermediate node. Such algorithms are more representative: directed diffusion(DD), greedy incremental tree(GIT), LEACH and TEEN that based on hierarchical routing and PEGASIS that based on the chain routing.From the perspective of network topology, the wireless sensor network routing protocols can be divided into two categories: plane and hiberarchy. It is equal status for each network nodes in plane routing protocols, and not a bottleneck effect, but the lack of management nodes to optimize communications resources, self-organization cooperate with other algorithm complexity, the dynamic response slow upon the network changes, and nodes near the sink die quickly because of the energy exhausted. The hiberarchy routing protocol can greatly reduce energy consumption overall network, and can enhance the network extendability.Majority of hiberarchy routing protocol based on the clustering routing protocol. Cluster-Based protocol designed to restrict a large amount of data communication and aggregation in the cluster (local), which reduces energy consumption. This paper aim at the issues that energy consumption is not balanced of different regions in existing clustering algorithm, propose an improved energy efficient even algorithm, REEEC.REEEC integrated consideration of the survival time, extendability and load balancing, has no special requirements for the distribution of nodes and capacity, is a energy efficient clustering algorithm.There is a great deal data aggregation in sensor networks,which only is interist in some statistical information of sense data, and does not need to acquire a large number of original sense data. For aggregation query, in-network aggregate operation can be used for specific optimization, and make full use of storage and computation capacity of local nodes, aggregate sense data on intermediate node of the aggregation tree, reduce the volume of data transmission.This paper integrate clustering algorithm,multihop routing with application layer data aggregation for proposed and implemented a cross layer data aggregation model, and try to quantify impact and effectiveness of clustering and data aggregation for network lifetime and the number of successful transmission of measurement data.A data aggregation tree is constructedfor in-network aggregation. The tree is rooted at the observer. which is the finaI destination of the aggregate report. Data aggregation reduces the communication overhead in the network. Thus saying the sensor scarce energy resourc. In addition, aggregation results in less channel contention and packet collisions. Conseuilently, data can reach its destination faster and more reliably.Prior to constructing the data aggregation tree. The network is c1ustered to identify a set of cluster heads that have higher average residual energy than their peers. 0nly cluster heads then proceed to discover the path to the root of the tree(the observer)by constructing a breadth-first spanning tree, Therefore, a cluster head acts as an aggregation point for its cluster members, as well as its child cluster heads in the data aggregation tree.For the view of the above, this paper propose a model and gives a concrete implement of the model and residual energy estimate method. Network compute to identify those with high residual energy cluster nodes as cluster head in each epoch. The cluster head constructed a spanning tree root of observer. Cluster head aggregate data from the cluster nodes and child cluster heads within the tree by simple aggregation operation, then the result of aggregation send along the tree to observer, thus reducing the volume and cost of communication and saving energy.Based on the experimental results, REEEC clustering algorithm select node those with high remaining energy as cluster head periodically, can balance energy consumption of nodes in the network. Query aggregation technology can aggregate data from cluster members or child cluster heads to effectively reduce the transmission network information, thus effectively prolonging the network lifetime, increased the success rate of data transmission. .