一种传感器多模态数据融合隐私保护方法  

A Privacy Protection Method for Sensor Multimodal Data Fusion

作　　者：于昕 率佳静 姜京文 YU Xin;SHUAI Jia-jing;JIANG Jing-wen(Special Education Research Center,Changchun University,Changchun 130022,China;School of Computer Science and Technology,Changchun University of Science and Technology,Changchun 130022,China)

机构地区：[1]长春大学特殊教育研究中心,吉林长春130022 [2]长春理工大学计算机科学技术学院,吉林长春130022

出　　处：《长春师范大学学报》2021年第12期59-64,共6页Journal of Changchun Normal University

摘　　要：无线传感器网络的数据融合隐私保护机制主要使用了同态加密的技术和基于信息交互的数据融合隐私保护技术,针对这两种技术中存在的对查询服务器(QS)不具备隐私保护性以及数据通信量太大的问题,提出了一种传感器多模态数据融合隐私保护方法——ESPART算法,该算法一方面通过分配随机时间片以躲避碰撞;另一方面为了减少数据通信量而选择依托有关数据融合的树型结构。与此同时,该算法对串通数据的范围进行控制管理,使数据丢失对精确度的影响尽可能地减小。根据仿真结果可以看出,与SMART算法相比,ESPART算法能够在数据隐私得到保障和数据通信量支出较少的情况下,得到更准确的数据融合结果。The data fusion privacy protection mechanism of wireless sensor network has the technology of homomorphic encryption based on information interaction.In order to solve the problems of lack of privacy protection for QS and large data traffic in these two technologies,a privacy protection method for sensor multimodal data fusion is proposed-ESPART algorithm.On the one hand,the algorithm can avoid collision by assigning random time slices;on the other hand,in order to reduce the data traffic,the tree structure based on data fusion is chosen.At the same time,the algorithm controls and manages the range of collusive data,so as to minimize the impact of data loss on accuracy.According to the simulation results,compared with SMART algorithm,ESPART algorithm can get more accurate data fusion results under the condition of data privacy protection and less data traffic expenditure.

关 键 词：ESPART算法 数据融合 隐私保护 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]