Research On Human-Computer Interaction Technology Based On Passive RFID

With the introduction of a series of concepts such as the Internet of Things and metaverse in recent years,more and more perceptual recognition technologies are being applied to human life.Virtual reality interaction is an important application area of perceptual recognition technology,which requires users to have an immersive experience when interacting with virtual environments,while traditional human-computer interaction devices such as mouse and keyboard cannot meet these requirements.The human-computer interaction devices based on perception recognition technology can sense the user’s body movements,so that the user can directly interact with the surrounding devices or virtual reality environment through his own body movements,which solves the shortage of traditional human-computer interaction devices in the field of virtual reality interaction.In this thesis,we discuss RFID-based Human-Computer interaction technology and introduces the current status of related research.Based on this,we design two kinds of passive RFID-based Human-Computer interaction methods: Passive RFID phase information based arm motion sensing method and Passive RFID phase information based person step orientation recognition method.In the Passive RFID phase information based arm motion sensing method,we first attach the passive RFID tag to the user’s arm and collect the phase data of the tag during the arm motion;then we perform data pre-processing on the collected phase data,including phase unwrapping,data smoothing and interpolation;then,by using the method of calculating the rotation angle of arm joints based on the tag phase data proposed in this thesis,the rotation angles of shoulder joints,elbow joints and palm orientation during arm movement are calculated by solving a system of equations,and by tracking the rotation angles of these joints,the tracing of arm movement trajectory is completed.In the Passive RFID phase information based person step orientation recognition method,we first collect the phase values of the user’s stepping movements in different directions through a set of 3X3 tag arrays;then pre-process the collected phase values,including phase unwrapping,data smoothing,interpolation,and normalization;then divide the movement data by the action division method based on the amplitude measurement values and frequency measurement values;finally,the CNN convolutional neural network is trained,and the recognition of the user’s stepping direction is performed by the CNN network.The experimental results show that the average recognition accuracy of this method can reach more than 95% for eight different directions of stepping movements.