Design And Implementation Of Identity Authentication Method Based On Mobile Phone Carrying Behavior Analysis

With the rapid development of smartphones,smartphones have gradually evolved into a smart housekeeper,which provide various services,making human life more convenient and faster.However,the issue of personal information security in smartphones has become increasingly prominent,as more and more significant and private information needs to be stored in mobile phones,and its importance has greatly exceeded this device.Therefore,the access control of smartphones has become the first line of defense to protect information security of mobile phone.Mobile identity authentication technology has also become a hot research topic.Most of the traditional authentication mechanisms for mobile phone are based on passwords or PIN codes,and they have various risks such as being vulnerable to brute force or stealing.Therefore,this paper makes full use of the data collected by the built-in sensors of mobile phones,and studies the biometric identity authentication method based on mobile phone behavior pattern recognition.Human behavior has a huge connection with innate genes and acquired habits.Therefore,users' behaviors are characterized by strong identification and identity authentication based on their behavior,which are continuous,invisible,inactive and freedom from the external environment.However,there are many random states in the process of carrying and using the smartphone,such as the direction of the mobile phone,the wearing position of the mobile phone and the user behavior.The sensor data collected in each state has a large difference as well.Therefore,in order to improve the user identification rate,this paper divides the identity authentication process based on carrying behavior pattern recognition of smartphones into two stages.First of all,the carrying behavior of smartphones is identified,wherein the carrying behavior mainly refers to the wearing position of the mobile phone and the behavior information of the user.In this paper,the multi-classification support vector machine is combined with the posterior probability calculation to obtain the classification posterior probability of each carrying behavior.Then,based on the time correlation of the user's carrying behavior,the carrying behavior recognition result is filtered to improve the carrier behavior recognition rate.Moreover,the identified smartphones carrying behavior is mainly two discrete features of user behavior and mobile phone wearing position.This paper encodes these two features by means of unique thermal coding,and then integrates these features into the user identity authentication feature for user identity,which generates to determine the identity of user.The experimental results show that the identity authentication based on carrier behavior recognition of smartphones achieves an accuracy rate of 91.6%,which is a great improvement compared to the authentication method without mobile phone carrier behavior recognition.