Research On Identity Identification Method Based On CSI

The 21 st century is a time of rapid development of science and technology,and the progress of computer technology has promoted the technological process of various industries.The ubiquity of intellectualization has made the task of manual operation in the past gradually replaced by machines,and various intelligent management systems are widely used in various fields.Intelligent development need to rely on intelligence system to provide security services,so no matter what scenario system privacy and security are user focus on content,to ensure the system security requires system can carry on the identification to the user,so the identification method of the research is of great importance.Biometric identification method has been widely used in the practical production and life,and get better effect,it can guarantee the high identification accuracy,one of the most representative is the fingerprint recognition,facial recognition and voice recognition.Although there is no denying the fact these identification way to have an advantage in reliability,but there are also some shortcomings,specific identification device is needed to identify users,and in the process of identification can be perceived by the user.In this paper,the method of identity identification using channel status information can be used to identify the indoor people under passive and unperceived conditions.The research focuses on two aspects.On the one hand,the references in the past line-of-sight path location when the channel state information,the most obvious fluctuation in this article in view of the current characteristics of the deficiency of interval locating method,put forward a kind of dynamic positioning feature interval selection algorithm.The characteristics of interval selection algorithm can accurately locate logo identity information of the channel state information range,another channel state information,which contain abundant message of subcarrier,also contain redundant information.In view of the rich channel state information,this article USES principal component analysis method,to choose the most sensitive in 30 is the carrier of subcarrier information identification for people to get the most effective information,the method of using machine learning to the person's identity identification classification,finally through the contrast experiment to prove the feasibility and effectiveness of the method.This paper summarizes the identification method based on channel status information under thefixed path condition and improves the identification rate.On the other hand,because there are many influencing factors in the non-fixed path conditions,the characteristics of identity identification are uncertain.In this paper,the identification method is used to study the identification method in the non-stationary path.Therefore,an identification method based on ReliefF algorithm is proposed.The method USES ReliefF feature selection algorithm,using the sorting after the characteristics of the combination of identity identification without a fixed path condition,and combined with Adaboost algorithm for multiple weak classifier for training,improve the effect of classification.Walking enables users to any position indoors is complete identification,precluding experimental process carried out by line-of-sight path location identification,the final validation method is feasible through the experiment.