Research On The Key Technology Of Continuous Authentication For Mobile Intelligent Terminal

With the continuous popularity of mobile devices,the use of mobile devices is more and more frequent,the security of mobile devices is widely concerned.Various authentication mechanisms have emerged to address the security and privacy of mobile devices.At present,the identity authentication mechanism of mobile devices belongs to explicit identity authentication,which is only for the single unlocking of mobile devices and does not guarantee the security of mobile devices after unlocking devices.In this thesis,the continuous authentication of mobile devices is taken as the research object.The continuous authentication based on motion sensor is improved,a multimodal fusion scheme based on motion sensor and face is proposed,and a complete multimodal continuous authentication process is constructed.The specific work is as follows:(1)Because most of the studies ignored the motion sensor data fluctuation caused by different mobile phone holding orientations,this thesis analyzes the reasons for the influence of mobile phone holding orientations on the motion sensor data change.Considering the influence of data on the accuracy of authentication in the case of multi-orientations,an adaptive continuous authentication scheme for mobile phone holding orientations is proposed.Experiments show that this method can better adapt to the scene of users using mobile devices in different mobile phone holding orientations and improve the accuracy of authentication.(2)Because single biometric may be interfered by noise data,non-universality,simulated impersonation and other factors,multiple biometrics authentication technologies are integrated to improve the security and reliability of the authentication system.Considering the computing power,processing speed and authentication accuracy of mobile devices,motion sensor and face were selected as the biometric features of the authentication scheme,and one-class classifier was used to train the data.The quality score of each biometric is calculated,and the quality score is combined with the matching score for the multimodal fusion.At the same time,the authentication scheme is implemented on the real machine,and the authentication experiment is carried out simultaneously on the open data set and the real machine.The experiment shows that the authentication accuracy of the scheme is98.53%.The authentication system can better adapt to the changing usage scenarios during the use of mobile devices.