Research On Gesture Recognition And Identity Authentication Technology Based On Channel State Information

In recent years,with the rapid development of communication networks,indoor human body recognition and detection have received great attention as key enabling factors for the construction of smart environments such as smart homes,smart retail,and smart museums.Compared with traditional vision-based and wearable sensor-based solutions,RF-based methods stand out because of their non-contact and non-line-of-sight perception.Among all RF-based methods,Wi Fi-based methods have become the focus of many researchers due to their ubiquitous availability and cost efficiency.The Channel State Information in the Wi Fi signal provides more fine-grained and richer information for the realization of gesture recognition and identity authentication.This lays a good foundation for the development of smart home and smart environment.First,to address the problem that the current gesture recognition technology based on Channel State Information can only recognize a user's single-facing gesture,a gesture recognition scheme based on Channel State Information,Wi-MG,is proposed.The method is based on the Channel State Information of different gestures,and each gesture contains data in four directions.Use wavelet analysis to remove the noise information in the data.The collected data usually contains part of the data without gestures.This part of the data will affect the subsequent recognition.In order to improve the efficiency and accuracy of subsequent training and classification,the improved the short-term energy algorithm divides the interval containing gesture actions.Finally,the mean value,standard deviation,correlation coefficient and average absolute deviation are extracted as features,and the Support Vector Machine classifier is used for classification,which realizes multi-directional gesture recognition.Secondly,to solve the problem of narrow application scope of the existing identity authentication method based on Channel State Information,a gesture recognition-based identity authentication method Wi-GI is proposed.This method uses principal component analysis to process experimental data to highlight individual differences in the same gesture;and by introducing frequency energy maps,the method of extracting features from images instead of extracting features from amplitude data,and then using convolutional nerves the process of network realization of identity authentication.Through experimental verification and analysis,the method proposed in this thesis has a high accuracy rate and achieves a good identity authentication effect.