无线传感网络中数据聚合隐私保护方案的研究

发布时间：2019-04-16 11:05

【摘要】：无线传感网络是一代新型的传感器网络,由若干廉价节点组成,负责采集和处理数据。无线传感网络可以大量感知信息(如地震数据、声音数据、温度、湿度等)进行分布式处理,提高感知数据的精确性。无线传感器网络深入我们生活的每个环节、渗透社会的每个角落,有利于帮助人类提高认识物理世界的深度、广度、精确性、及时性,加强和密切人类与物理世界的联系,大力提高人类对物理环境的远端监视和控制能力,所以无线传感器网络有着非常广阔的应用前景。随着无线传感网络的快速发展,它的安全问题受到了广泛的关注,本文对无线传感网络中的数据聚合隐私保护方案进行了研究。根据目前关于无线传感网络的研究,数据聚合隐私保护的方案主要有基于数据扰动的隐私保护、基于安全多方计算的隐私保护、基于同态加密的隐私保护、基于多项式回归的隐私保护。本文重点基于多项式回归的隐私保护和基于同态加密的隐私保护方法提出新方案。本文的主要贡献有:第一,本文提出PRDA+协议。首先,节点将n个传感数据通过最小二乘法拟合成多项式函数,并将多项式函数的系数代替原始节点数据上传至聚合器端,大大降低了通信量。其次,我们的方案中,节点与基站会预先部署私密密钥,随机数生成器通过私密密钥为种子生成随机数进行加密,较好地保护了数据的隐私性。最后,我们通过二元数据之间的关联性将解密后的聚合数据进行对比,实现了数据的完整性保护。第二,我们提出了可进行验证的加权平均聚合方案(VAAS),针对用户信息重要程度各不相同,我们利用权重的概念,为每个传感器节点设置相对应的权重。信息传输过程中,如果单纯采用加密来保证数据的隐私性,实际上还存在缺陷,假如发送方在发送某个信息后突然反悔,声明这个信息不是其发送的,进行抵赖,虽然数据传输过程中是保密的,但无法证明这个信息的发送方身份,这样不利于通信的管理。针对这个问题,我们采用的是ElGamal数字签名对传感器节点进行身份验证。在聚合器进行聚合运算之前,需要先对节点的数字签名进行验证。签名验证算法使方案具有抵抗数据篡改和数据抵赖、追查错误数据来源的能力。
[Abstract]:Wireless sensor network (WSN) is a new generation sensor network, which is composed of several cheap nodes and is responsible for collecting and processing data. Wireless sensor networks (WSNs) can process a large number of sensing information (such as seismic data, sound data, temperature, humidity, etc.) to improve the accuracy of sensing data. Wireless sensor networks penetrate every link of our life and penetrate into every corner of society. It is helpful to help human beings improve their understanding of the physical world in depth, breadth, accuracy, timeliness, and strengthen and close the connection between human beings and the physical world. Wireless sensor networks (WSNs) have a wide range of applications due to the improvement of the remote monitoring and control ability of human beings to the physical environment. With the rapid development of wireless sensor networks, its security issues have received extensive attention. In this paper, the privacy protection scheme of data aggregation in wireless sensor networks is studied. According to the current research on wireless sensor networks, data aggregation privacy protection schemes mainly include privacy protection based on data disturbance, privacy protection based on secure multi-party computing, privacy protection based on homomorphic encryption, Privacy protection based on polynomial regression. This paper focuses on privacy protection based on polynomial regression and privacy protection based on homomorphic encryption. The main contributions of this paper are as follows: first, the PRDA protocol is proposed in this paper. Firstly, n sensor data are fitted into polynomial function by least square method, and the coefficients of polynomial function are uploaded to the aggregator terminal instead of the original node data, which greatly reduces the traffic. Secondly, in our scheme, nodes and base stations will pre-deploy private key, and random number generator can generate random number to encrypt by seed of private key, which can protect the privacy of data better. Finally, through the correlation between binary data, we compare the decrypted aggregate data and realize the integrity protection of the data. Second, we propose a verifiable weighted average aggregation scheme, (VAAS), which is different in the importance of user information. We use the concept of weights to set the corresponding weights for each sensor node. In the process of information transmission, if only encryption is used to ensure the privacy of the data, in fact there is still a defect. If the sender suddenly reverses after sending a certain information and declares that the information is not sent by it, it is repudiated. Although the data transmission process is confidential, but the identity of the sender of this information can not be proved, which is not conducive to the management of communications. In order to solve this problem, we use ElGamal digital signature to authenticate the sensor node. Before aggregating the aggregator, the digital signature of the node needs to be verified. Signature verification algorithm has the ability to resist data tampering and data repudiation, and to trace the wrong data sources.
【学位授予单位】：浙江工商大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP212.9;TN929.5

【相似文献】