The Research Of User’s Authentication System Based On Rhythm Feature

To solve the problem of portable device user identification, this article designed a kind of user identity authentication system based on the characteristics of rhythm. The system collected rhythm information of the key sequences. With the combination of adaptive clustering algorithm and time threshold algorithm, the rhythm characteristic code characterizing user identity information was extracted to form the rhythm of the key. Eventually, the user authentication was realized. This system is based on SQL server 2012 and Visual Studio 2012.The article contrastively used two kinds of user rhythm recognition methods. In the early time, it matched the rhythm of the key with the proportion algorithm. Based on the shortest time interval in the users’ rhythm time interval sequence, other time interval was divided by it. We converted the ratio to an integer considering the unstable state of users’ rhythm. Thus, the user identification code of rhythm was formed. Found in the test, although the scheme could preliminarily identify user identity, the recognition was not high. The test data indicates that this method of intuitive extraction of user rhythm feature could not effectively identify the user identity. In the later work, considering the problem that proportion algorithm could not accurately classify user’s key sequences. So adaptive clustering algorithm was used to quickly classify the time series. Taking the destabilizing factors of the user rhythm into account at the same time, the time threshold algorithm was introduced into adaptive clustering algorithm. Thus it could exactly determine the changes of two adjacent rhythm of time interval and extract the user characteristics of rhythm and form the corresponding rhythm identification code.The test which was made by 19 users verified the validity and intelligence of the adaptive clustering algorithm and time threshold algorithm in the article. The test data shows that the algorithms used in this article could match the rhythm of the key well and effectively identify user identity information.