行走状态下基于人体通信的身份识别方法研究  

Research on identity recognition method based on human body communication in walking state

作　　者：吴秋雯 廖薇 WU Qiuwen;LIAO Wei(School of Electrical and Electronic Engineering,Shanghai University of Engineering Science,Shanghai 201620,China)

机构地区：[1]上海工程技术大学电子电气工程学院,上海201620

出　　处：《武汉大学学报(工学版)》2025年第3期504-512,共9页Engineering Journal of Wuhan University

基　　金：国家自然科学基金资助项目(编号:62001282)。

摘　　要：针对静态下的人体生物特征身份识别使用数据较为单一的问题,提出一种行走状态下基于人体通信的身份识别方法。首先,通过时域有限差分法(finite difference time domain method,FDTD)建立多个行走状态下的人体模型,并设置多条通信链路以获取原始数据集;其次,将原始数据集进行分段聚合近似处理;最后,将其输入随机森林(random forest,RF)、KNN(k-nearest-neighbors)、极端梯度提升(extreme gradient boosting,XGBT)以及融合模型进行训练,从而通过测试获得最终识别效果。结果表明,利用行走状态下获取的路径衰减进行身份识别具有较好的效果,为人体通信在身份识别领域的应用提供了新思路;同时,融合模型在数据集上得到的准确率为97.05%,较之单一模型得到了明显的提高,且处理后的数据集在训练时可以有效提高速度,实现快速识别。Aiming at the problem that the data used for human biometric identification in static state is relatively simple,an identity identification method based on human body communication in walking state is proposed.Firstly,finite difference time domain method(FDTD)is used to establish multiple human body models in walking states,and multiple communication links are set up to obtain the original data set.Secondly,the original data set is approximated by segmented aggregation.Finally,it is fed into random forest(RF),Knearest-neighbors(KNN),extreme gradient boosting(XGBT),and fusion model for training to obtain the final recognition effects through testing.The results show that the use of the path attenuation obtained in walking states has a good effect on identity identification,which provides a new idea for the application of human body communication in the field of identity identification.At the same time,the accuracy of the fusion model on the data set is 97.05%,which is significantly improved compared with the single model,and the processed data set can effectively improve the speed during training to achieve fast recognition.

关 键 词：路径衰减 人体通信 生物识别 动态模型 

分 类 号：TN915[电子电信—通信与信息系统]