Algorithm For Wireless Sensor Data Networks

Wireless Sensor Network（WSN） is developing rapidly as a new type of intelligentmonitoring network. The micro sensors of wireless sensor network supervise and gatherinformation by cooperative work. Then they take the information processing efficientlythrough the embedded computing systems. They are formed in dynamic networktopology with the self-organizing manner, through wireless communication transmissionto the base station. The network usually deployed in harsh environment monitoring areamonitoring information acquisition task, and due to its own characteristics, the sensornode energy is limited, the energy consumption in the network are mainly concentratedin the wireless communication transmission phase. Using data fusion technology caneffectively reduce the redundant information and error information, improve thetransmission efficiency of nodes, thereby saving the node energy consumption andmaking the network life cycle extend.For this purpose,many academics multi-angle analysis, make data fusion technologyof the wireless sensor network in the network layer and application layer becomes a hottopic concerned research. On the basis of previous studies, I will combine the WSNnetwork layer routing protocol with the data fusion algorithm based on BP neuralnetwork in this paper, and I put forward a new data fusion algorithm based on BP neuralnetwork is DCBP. The main research work and innovations as follows:1. The study of wireless sensor network and data fusion technology are researchedand analyzed the research status and related knowledge, according to the shortcomings ofthe existing WSN network layer clustering routing protocol proposes a doublecluster-heads routing protocol based on polymerization for wireless sensornetworks named DCHP. In this paper we analyzed the low energy adaptive clusteringrouting algorithm LEACH and clustering routing algorithm based on energy predictionCHEP, because of the LEACH algorithm does not take into account the residual energyproblem and the uneven distribution of the cluster heads,CHEP algorithm does notconsider the spatial location of cluster head nodes and the node’s data correlation, whichimpacted the data fusion. The agreement from the node degree of polymerization,residual energy and the node space position, etc was improved, that keep the balance of network energy consumption and improve the network performance.2. Neural network and BP neural network are studied, and according to the WSNdata fusion technology and the BP neural network function similarity, I put forward adata fusion algorithm based on BP neural network named DCBP. Analysis of neuralnetwork and BP neural network, because of the WSN data fusion technology and the BPneural network can be calculated by certain rules to deal with a large number of data andinformation, so as to get the results reflected the characteristics of the data. Therefore theBP neural network algorithm is introduced into the improved wireless sensor networkrouting protocol DCHP. Three layer BP neural network is applied to each clustering,within the cluster nodes as input layer neurons in preliminary data processing; And thenthe results to the first cluster heads, and by the first cluster heads according to neurons inhidden layer and output layer neurons function further processing; Finally the adjacentsecond cluster head with the feature values through multiple hops routing sent to the sinknode. The new data fusion model dealing with the amount of data is much less than theamount of the source data collected, thus it reduces the energy consumption in datacommunication, prolong the network life cycle effectively.3. The data fusion algorithm DCBP is simulated by NS-2network simulationplatform, respectively from the number of node survival, total energy consumption andthe relationship between the average node energy consumption and node number threeaspects in WSN on the analysis comparison, that it verifies the validity of the DCBPalgorithm, improve the network performance.