基于一维卷积神经网络的物联网终端非干扰身份识别方法  

Non-Interference Identification Method of IoT Terminals Based on One-Dimensional Convolution Neural Network

作　　者：杨萌[1] YANG Meng(Department of Computer Science and Technology,Huaibei Vocational and Technical College,Huaibei 235000,Anhui China)

机构地区：[1]淮北职业技术学院计算机科学技术系,安徽淮北235000

出　　处：《吉首大学学报（自然科学版）》2024年第4期26-31,共6页Journal of Jishou University(Natural Sciences Edition)

基　　金：安徽省高校自然科学研究重点项目(KJ2021A1373,KJ2021A1376,KJ2020A0964);淮北职业技术学院自然科学研究重点项目(2020-A-5)。

摘　　要：针对物联网终端身份识别过程易受干扰的问题,设计了基于一维卷积神经网络的物联网终端非干扰身份识别方法.通过物联网终端内置的加速度传感器收集不同步态的加速度数据,并采用平滑与分窗处理进行预处理.利用步态数据融合模型分析用户在不同行为位置组合下的数据,再利用随机森林模型预测物联网终端的位置与用户行为.将行为与位置数据融合,得到步态融合数据,再将融合数据输入至构建的一维卷积神经网络模型中,经过迭代训练后输出用户的身份识别结果.实验结果显示,基于一维卷积神经网络的物联网终端非干扰身份识别方法在合适的模型参数条件下能够准确识别物联网终端用户身份.In view of the problem that IoT terminals are vulnerable to interference in the identification process,a non-interference identification method of IoT terminals based on one-dimensional convolution neural network is proposed.Acceleration data for different gaits are collected through the built-in acceleration sensor of the Internet of Things terminal and preprocessed through smoothing and windowing processing.Gait data fusion model is used to analyze user data under different behavior-position combinations,and random forest model is used to predict the position of IoT terminals and target data behavior.Behavioral target data are integrated with positional data to obtain gait fusion data.On this basis,a onedimensional convolutional neural network model is constructed,which takes the fused data as input and outputs the user's identity recognition results after iterative training.The experimental results show that the non-interference identity recognition method for IoT terminals based on one-dimensional convolutional neural networks can accurately identify the identity of IoT terminal users in appropriate model parameter conditions.

关 键 词：卷积神经网络 物联网终端 非干扰 身份识别 步态数据 行为位置组合 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]