Research And Application Of Smart Phone User Authentication Algorithm

With the rapid development of science and technology,MobileInternet has penetrated into all aspects of life,and smart phones have gradually become an important carrier for information transmission and exchange in the mobile Internet.The smart phone stores a large amount of personal information such as contact lists,call records,mobile phone short messages,personal photos and so on.Once these data are leaked or used,users' data privacy will be stolen and even affect national security.Generally speaking,mobile phone users must first perform user authentication.The current general authentication methods are divided into traditional password authentication and biometric-based user authentication.However,these authentication methods can be broken through prying,theft,or some high-tech means,which seriously endanger personal privacy.Therefore,there are many disadvantages in the current mobile phone user authentication method.The smart phone always records the daily behavior of users and each person's behaviors are also different.After collecting these data for further analysis and processing,we can determine whether user authentication can be performed to avoid user privacy leakage.This is also the research direction and purpose of this article.The main research contents of this article are as follows:1.This paper summarizes the current research status of user authentication at home and abroad,analyzes the advantages and disadvantages of commonly used mobile phone user authentication methods,and proposes a user authentication method based on the mobile phone user's behavioral characteristics during the use of the mobile phone.The authentication method can detect that the attacker's behavior is different from that of the user when the user forgets to lock the screen,and locks the mobile phone.As a supplement to the traditional user authentication method,this further protects user information security.2.According to the collection object,the type of data collected,the acquisition method,and the data storage method,we designed a user behavior data collection system and deploy it on the Android system.At the same time,we perform long-term cycle collection of behavioral data for different users and migrate to the database via Python scripts to facilitate post-processing analysis.3.The user behavior data collection system will clean,integrate,denoise the collected data and make other kinds of data preprocessing.The system also makes various statistical characteristics of network traffic information and sensor data to form a time series.The DTW algorithm is used to differentiate the users to calculate thesimilarity between two different users or the same user.It can be clearly determined whether the owner is in the process of using the mobile phone.4.For the behavior analysis of different users when answering a phone call,we found that due to the shape of the upper limb,the length of the arm,and the user's arm strength,the sensor data will constantly change from picking up the phone to answering the call.The system will extract the mean value variance,maximum and minimum values,peaks and valleys,and other characteristic values of the sensor data,and classify the users by the SVM algorithm to accurately complete the user authentication.Compared with Bayesian algorithm and decision tree classification algorithm,experiments show that the SVM algorithm has the best user authentication and achieves85.9% accuracy.