The Research Of Wireless Sensor Networks Intelligent Clustering And Data Fusion Algorithm

Wireless sensor networks (WSN) is a burgeoning next-generation network. It is deemedas one of the most influential technologies in the21stcentury which could change the world.WSN combines many areas,such as embedded processing technology, distributed information,wireless communication technology and so on. With the development of these technologiesmakes the WSN have a wide application apace, for example military the applications in thefield of defense, environmental monitoring, public health and so on. It is composed of a largenumber of sensor nodes which are distributed in the testing area. These nodes send theinformation which is collected from the area to the sink node,then the sink node manages thedata and evaluates testing area.Since2004, Candés, Romberg and Tao put forward Compress Sensing(CS) theory andthis theory officially becomes independent theory. CS theory bring great changes in signalprocessing just ten years, This theory was firstly used in the study of nuclear magneticresonance (NMR) imaging. It was applied to wireless sensor network (WSN), engineeringmathematics and image processing and other disciplines. At present, the CS development isvery rapidly. Headed by the United States, the famous universities of other country havebegun to research the related theory of CS. The premise of the signal can be compressed if thesignal can be sparse, so that the sampling and compression for the signal can be processed atthe same time. Measurement matrix projection can change high-dimensional signal to the lowvisa signal, which can reduce the amount of signal transmission and the network energyconsumption, and can solve the problems of signal transmission congestion. The low visasignal high-dimensional signal can be reconstructed by reconstruction algorithm. In view ofthe node energy limited, this article proposed an algorithm which is hierarchical data fusionbased on distributed compressed sensing:Combining with clustering protocols and nodeinformation relevance, using Joint Sparsity Model(JSM) reconstruction algorithm toaccurately reconstruct of node information, experiments prove that this algorithm can savenode energy consumption and improve the accuracy of signal reconstruction. In this paper, themain research work is as follows.(1) In the traditional theory of CS, each WSN sensor node sends the information to sinknode directly, the internal correlation of node information is only considered when it isreconstructed, which may lead to the loss of the node information and too much consumenetwork energy. In order to adjust the processing performance and limit WSN node energy,this paper proposes an improved algorithm of distributed compressed sensing based on edge information. This algorithm combines with LEACH clustering protocols, and utilizes thespatial correlation of wireless sensor network node with JSM1sparse model. In addition, thesignal information is reconstructed according to the distributed compressed sensing algorithmbased on side information.(2) This article in view of the the disadvantages of LEACH algorithm int the wirelesssensor network clustering:clustering is not uniform, easy to make a node as the cluster headrepeated,so the sensor energy consumption early finish for the dead sensor,the networkappears a blind spot to collect data.The traditional artificial fish with slow convergence speedin the sensor network and large amount of calculation.On the basis of LEACH and traditionalartificial fish algorithm for network clustering dynamic change crowded degree, artificial fishprogress before long and visual field in each iteration making the web from the global searchand local search.To accelerate the convergence speed and reduce11%of the amount ofcalculation, save21%of the network energy, so as to prolong the network lifetime(3) If network meet JSM2sparse model, put forward another algorithm, which isdistributed compressed sensing based on hierarchical data fusion: This paper adopts thewireless sensor network clustering algorithm based on the dynamic artificial fish optimizationalgorithm. The node data are transmitted to their cluster heads respectively and in the nodesdata are encoded independently the cluster head, then the measured value sent to the SINKnode, and the data of the entire network node are reconstructed by SOMP reconstructionalgorithm in the SINK node. The simulation results of algorithm can prove that thisalgorithm compared with LEACH, AFSO this algorithm can effectively reduce11%of thenetwork energy consumption and21%of prolong the network life cycle. Adoptingimproved distributed compressed sensing algorithm can not only obtain accurately reconstructvalue of the node, but also can greatly reduce the energy consumption in the process ofalgorithm realization.