基于BP网络判断传感器数据可信度研究  

Research on data credibility of sensor based on BP network

作　　者：刘小久 袁丁[1] 梁瑷云 严清[1] Liu Xiaojiu;Yuan Ding;Liang Aiyun;Yan Qing(School of Computer Science,Sichuan Normal University,Chengdu 610101,China)

机构地区：[1]四川师范大学计算机科学学院

出　　处：《计算机应用研究》2019年第8期2440-2443,2453,共5页Application Research of Computers

基　　金：国家科技支撑计划资助项目(2014BAH11F01);国家自然科学基金资助项目(61373163);四川省可视化与虚拟现实软件重点实验室项目

摘　　要：传统方法使用对称及非对称加密对传感器网络系统进行安全保障,需要大量的加/解密计算且在密钥被破解后不能准确判断数据的可信性,不能有效保证无线传感器网络系统安全。为保障无线传感器网络系统安全,针对无线传感器网络中节点信息可信度问题,提出了一种基于BP网络判断节点信息可信度的方法。该方法在边界路由器上使用BP神经网络,对采集的多特征值数据进行训练,然后用训练所得结果判断节点可信度,进而筛选出数据。该方法具有较低的系统开销与较高的安全保证,能够筛选出问题节点,并保证传感器网络的安全运行。实验结果表明,该方法认证时间短,能达到预期效果。The traditional method uses symmetric encryption and asymmetric encryption to secure the sensor network system, which requires a lot of encryption and decryption calculations and the credibility of the data cannot be accurately judged after the key cracked. The encryption method hardly guarantees the safety of wireless sensor network system effectively. This paper proposed a method determining the reliability of node information based on BP network to ensure the security of wireless sensor network system by the credibility of node information in wireless sensor networks. The method used BP neural network on the border router to train the collected data of multiple eigenvalues and determined the credibility of the node with training results to filter the data. The BP network judgment method could achieve lower system overhead and higher security assurance without additional communication data or calculation tasks. What’s more, this method could screen problem nodes to ensure the safe operation of the sensor network. Experiment confirms that this method with a short authentication time cost can achieve the desired results.

关 键 词：BP网络 无线传感器网络 传感器节点 可信性 安全性 

分 类 号：TP309.2[自动化与计算机技术—计算机系统结构]