全文下载：2013030309
蔡宗吟1， 刘才铭1， 刘毅1， 叶秋冬2
（1.乐山师范学院 智能信息处理及应用实验室,四川乐山 614000; 2.重庆理工大学 计算机应用技术研究所，重庆 400054）
摘要： 无线传感器网络对信息的安全性和可靠性比较敏感，在数据融合过程中易受到节点生存期短和数据被伪造或篡改等方面的威胁，为了提高融合结果的真实性与可靠性，本文提出一种针对网络路由和节点的优化改进策略，路由方面应用流量分布加权算法最大化网络的生命周期；节点方面，应用BP神经网络算法融合簇头节点和基站的数据，最后采用信誉度评价机制对网络的整体性能进行分析。实验表明，改进后的融合方法可靠性提高，较大程度减少了数据的冗余度、降低了能量的消耗，提高了整个网络的性能。
关键词： 无线传感器网络；数据融合；路由算法；BP神经网络；信誉度
中图分类号： TP 393文献标志码： A
Application of High Performance Data Fusion Algorithm in Wireless Sensor Network
CAI Zong-yin1， LIU Cai-ming1， LIU Yi1， YE Qiu-dong2
(1.Intelligent Information Processing and Applications Laboratory,LeshanNormalUniversity, Leshan 614000,China;
2.Institute of Computer Application Technology,ChongqingUniversityof Technology,Chongqing400054,China)
Abstract: The wireless sensor network is sensitive to the information safety and reliability and is vulnerable for node short survival and forged or tampered data in data fusion process. In order to improve the authenticity and reliability of the fusion results, we proposed in this paper an improved and optimized strategy of network routing and node. Flow distribution weighting algorithms are employed to maximize the network lifetime for routing, and BP neural network algorithm is used to integrate the cluster head node and the base station data for node treatment, finally the reputation evaluation mechanism is employed to analyze the network overall performance. Experimental result shows the improved fusion method enhances the reliability, decreases data redundancy in greater degree, reduces energy consumption and therefore improves overall network performance.
Key words:WSN; data fusion; routing algorithm; BP neural network; credibility
收稿日期:2012-09-25
基金项目: 国家自然科学基金项目（61103249）；四川省重点科技项目（2010JY0146）；四川省教育厅科研项目（12ZB239）；乐山师范学院科研创新团队建设计划项目（0300030）.
作者简介: 蔡宗吟(1980—)，女，讲师.
