Research On Human Gesture Recognition Based On Boost Method In Wi-Fi Environment

With the development of science and technology,people's lives are becoming modernized and intelligent gradually.Human-computer interaction is the key to realize the smart life,and the human gesture recognition is an important means of human-computer interaction.As the most ubiquitous wireless signal in life,Wi-Fi is of great significance in the field of human-computer interaction.In this paper,the identification method of human gesture in Wi-Fi environment is studied.By using signal processing and pattern recognition technology,the method of improving recognition accuracy and recognition speed is explored.In order to extract effective signal characteristics and improve the accuracy of human gesture recognition,this paper analyzes the influence of communication environment on signal transmission,and confirms the random fluctuation of amplitude envelope of the received signal caused by multipath transmission,which contains different dynamic human gestures information.Thus,a feature vector consisting of attenuation information at different degrees relative to the peak value of the received signal is proposed.In addition,the method of signal feature extraction and feature selection is discussed.The XGBoost method is used to sort the feature importance,and it is proved that the features extracted in this paper play an important role in the result of human gesture recognition.In this paper,a human posture recognition model based on ensemble learning method is proposed,which can take into account both the accuracy and speed of recognition.Compared with the traditional single classifier,the ensemble learning method combines the ability of multiple classifiers at the same time and gives a comprehensive discriminant result in order to obtain a higher recognition rate.The Adaboost gesture recognition method based on decision stump is able to reduce the limitation of parameter adjustment in different application scenarios while ensuring the recognition rate.The recognition method based on XGBoost can realize the feature selection in the process of model training,and ensure high recognition accuracy in the case of low dimensional eigenvectors.The collection of dynamic human posture data in Wi-Fi environment is based on the software radio platform.The proposed method is verified,and the experimental results show that the above method can obtain the expected recognition effect.The performance compared with the related algorithms shows the advantages of the proposed method in the recognition rate and speed.