基于神经网络的WSN数据融合改进算法研究

发布时间：2018-02-25 06:02

本文关键词： 无线传感网络数据融合神经网络模糊神经网络网络学习效率　出处：《太原理工大学》2014年硕士论文　论文类型：学位论文

【摘要】：无线传感器网络是由若干耗能较低,功能各异的传感器节点组成的。它们可以在不同的环境中监测和采集周边环境信息并将信息发送给工作人员。在此过程中,节点具有信息采集,处理和存储等功能,但考虑到其能源有限,且主要依靠无法替换的电池供电,同时采集到的信息具有高冗余性等特点,若是将这些数据全部发送给汇聚节点(Sink),会使节点能耗过快,降低网络使用效率。为了避免上述问题的产生,人们提出了数据融合(data fusion或data aggregation)技术。把数据融合应用于无线传感网络中,用以减少无线传感器网络的通信量,提高信息的融合度和准确度成为降低节点能耗、延长网络生命周期的主要手段之一。本文以环境监测为背景,首先提出了一种适用于WSN的基于神经网络的分簇路由协议数据融合模型。该算法将无线传感网络(WSN)的分簇路由协议与BP神经网络相结合,通过神经网络方法对簇内节点采集到的信息进行数据拟合,在此基础上,通过对网络训练参数的改进,网络训练收敛加快,缩短了网络收敛时长。最后,通过只将数据的特征值发送给汇聚(Sink)节点,以此来减少节点数据流量、节约能耗。通过仿真实验验证,该算法可有效减少网络通信量,降低节点能耗,延长网络寿命,同时还验证了本算法在环境监测等方面的实时性和有效性。再引入以T-S推理系统为基础的模糊神经网络数据融合方法,通过对模糊神经网络学习算法的学习、研究,提出了一种新的改进学习算法,最后再与分簇路由协议相结合,利用上文中提出的创新结合,提出了一种新的基于模糊神经网络的WSN数据融合模型。最后通过仿真实验表明,以水环境监测系统为背景,与传统的T-S模糊神经网络相对比,分别从网络预测准确度及网络收敛速率两方面,验证了改进算法模型的高效性,最终达到节省了节点能耗,延长网络寿命的目的,同时证明了其在水环境监测系统上的可行性及高效性。
[Abstract]:Wireless sensor networks are made up of a number of sensor nodes with low energy consumption and various functions. They can monitor, collect and send information about the surrounding environment to staff in different environments. The node has the functions of information collection, processing and storage, but considering that its energy is limited, and it mainly depends on the battery that can not be replaced, the information collected has the characteristics of high redundancy and so on. If all these data are sent to the convergent node, it will make the node consume energy too fast and reduce the efficiency of the network. Data fusion data fusion or data aggregation technology is proposed to reduce node energy consumption by using data fusion in wireless sensor networks to reduce the traffic of wireless sensor networks and improve the fusion degree and accuracy of information. One of the main means of prolonging the network life cycle. In this paper, based on the background of environmental monitoring, a clustering routing protocol data fusion model based on neural network for WSN is proposed, which combines the clustering routing protocol of wireless sensor network (WSN) with BP neural network. The neural network method is used to fit the information collected by the nodes in the cluster. On this basis, the network training convergence is accelerated and the network convergence time is shortened by improving the network training parameters. In order to reduce the data flow and save energy consumption, the algorithm can effectively reduce the network traffic, reduce the energy consumption and prolong the network lifetime by sending only the eigenvalues of the data to the convergent Sink node, and the simulation results show that the proposed algorithm can effectively reduce the network traffic, reduce the energy consumption of the nodes and prolong the network life. At the same time, the real-time and effectiveness of this algorithm in environmental monitoring is also verified. Then the data fusion method of fuzzy neural network based on T-S reasoning system is introduced. By studying the learning algorithm of fuzzy neural network, a new improved learning algorithm is proposed. Finally, it combines with clustering routing protocol. A new WSN data fusion model based on fuzzy neural network is proposed. Finally, the simulation results show that the improved algorithm model is more efficient than the traditional T-S fuzzy neural network in terms of network prediction accuracy and network convergence rate. Finally, the node energy consumption is saved and the network life is prolonged. At the same time, it is proved to be feasible and efficient in the water environment monitoring system.
【学位授予单位】：太原理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP212.9;TN929.5

【参考文献】