Fitness Motion Sensing Technology Based On WiFi Channel State Information

With the rapid development of pervasive computing,Internet of things and other technologies,the physical world and the information world are gradually merging to form an intelligent environment,and the premise of providing people-oriented intelligent application is the non-contact perception of human behavior.In a variety of sensing technologies,non-contact sensing based on WiFi signals can reuse existing communication infrastructure,which is not limited by light,line-of-sight,privacy protection,etc.,and has good universality and scalability.Moreover,as fine-grained WiFi channel state information can be obtained on commercial devices,the perception of WiFi channel state information has become a hot research,providing support for innovative applications in the Internet of things world.At the same time,there is a growing focus on personal health,with more and more people exercising at home to strengthen their bodies,build muscle,reduce the risk of chronic disease,and save time and money on going to the gym.Wearables,such as smartwatches,are often used for fine-grained recording and evaluation of fitness,but this contact-based sensing puts an additional burden on exercise.The contactless sensing technology in the intelligent environment is an ideal solution,but the use of WiFi signals for fine-grained recording and evaluation of fitness movements is still faced with the dilemma of not being able to analyze movements in detail and requiring people to be in the linear distance of sending and receiving.Aiming at the problems existing in existing studies,this paper proposes a fitness movement recognition scheme based on WiFi channel state perception to help people exercise at home more effectively.The main research results of this paper include:(1)A fitness action recognition method based on WiFi channel state information is proposed.In the transceiver linear distance WiFi channel state information of rapid attenuation of amplitude and phase noise change with time,this article clever use WiFi multiple Input multiple Output(Multi Input Multi Output,MIMO)communication mechanism,channel state information of the phase difference between the antenna as the original signal,segment the individual fitness movement after preprocessing,then using convolution neural network to extract the feature,realized the accurate identification of the fitness movement.(2)Put forward a fitness action quality assessment method based on WiFi channel state information.On the basis of fitness movement recognition,the convolutional neural network is used to recognize users.From the time dimension also define faster than,consistency,such as evaluation index,and based on the timing of the cumulative power spectral density estimation methods,segmentation method based on weighted dynamic time neat and a series of algorithm to obtain accurate time information for each of the fitness movement stages,implements the effective evaluation on the quality of the fitness movement.(3)A fitness assistance system based on WiFi channel state perception is implemented to help users exercise more effectively by giving them accurate feedback.The scheme of this paper was comprehensively evaluated,including motion detection,motion segmentation,motion recognition,user identification and motion time evaluation,etc.A large number of evaluation experiments verified the accuracy and effectiveness of the system in motion recognition,user identification and motion evaluation.This paper first introduces the WiFi based channel state information fitness movement perception of background and research significance,the analysis of the existing based on WiFi channel state information of the motion perception research progress and existing problems,and then put forward in this paper,the technical route,after detailed introduction based on WiFi channel state information of fitness action identification and quality evaluation method,design and implements a fitness auxiliary system based on WiFi channel state information and to do a lot of experiments on the proposed method is verified and evaluated.Finally,this paper summarizes the research content and looks.