Research On Key Technology Of Identity Recognition Based On Wireless Signal

With the development of society,identity identification has been closely related to our life.Authentication is required in many places of life,whether it's through a ID or password,fingerprint,face or iris,and all of these meth-ods have their limitations,such as the need to carry them with you or touch a human body,and the risk of privacy.Therefore,deviceless and contactless identification based on wireless signals has attracted extensive attention from researchers.Among them,with the widespread popularity of WiFi devices,identity recognition based on WiFi signals has become a research hotspot in recent years.This research mainly studies identity recognition based on WiFi signal from two aspects.On the one hand,the gait-based identification method is studied.Gait,as one of the biological characteristics of human beings,is widely used in identification.When people are walking,Channel State Information(CSI)in the surrounding WiFi signals will be affected.Through the analysis of CSI time series,the gait characteristics related to human identity information can be extracted,which can realize the identity recognition.However,the tradi?tional machine learning methods have insufficient generalization ability for the data collected at different times.Therefore,the data collected at different times are treated as source domain and target domain respectively,and the different domains are mapped to each other through the domain adaptive neural network model to improve the generalization ability.On the other-hand is the activity-based identification method research.People's activities also have a specific impact on CSI,from which individual specific information can be extracted to realize identity recognition.This re-search proposes a WAID(WiFi and Activity Based Identification)framework.Since the CSI waveform of different people doing the same activity is relatively similar,fewer features can be extracted in the time domain and frequency do-main.WAID uses two-dimensional fractional Fourier transform to transform the data into the fractional domain,so that more feature information can be extracted.The experimental results show that WAID has a good recognition effect in a variety of situations,and the average recognition accuracy of 97.7%is obtained in the experimental condition of 6 people.