Human Motion Recognition And Interaction Based On Smart Wearable Devices

With the advancement of technology and the continuous improvement of living standards,smart wearable devices have entered people's eyes and gradually integrated into people's lives,becoming a series of household electronic products.However,due to the wearable and portable characteristics of smart wearable devices,their size design is usually small,and they can only be equipped with a small touch screen or even no touch screen.As a result,the tap control that was originally suitable for mobile phones is difficult to use in wearable devices.Therefore,in order to improve the user's experience when controlling the wearable device,the included researchers are devoted to studying the human body's behaviors recognized by various devices to understand the intentions expressed by human actions.As an effective device control mechanism,human move-ments in human-machine interaction in existing research,either use additional special equipment to collect and analyze the corresponding data,or equip a finger or wrist with specific sensors,and so on.These methods are not suitable for daily human-computer interaction scenarios,and it also wastes a lot of power of wearable devices that are al-ready limited.In this paper,we will make full use of the category sensors(microphone,accel-eration sensor,gyroscope,and compass)embedded in smart wearable devices to es-tablish two systems to recognize human movements.First,for the interaction of the smart wearable device itself,large-scale microscopic movements,we propose a sound-based sensing system that utilizes fingers to perform gesture on the skin of the back of the hand.Compared with traditional interaction methods,it is more convenient,user-friendly and at your fingertips.Unlike previous work using dedicated devices for gesture input,our system uses passive acoustic sensing to recognize gestures,such as swiping to the left,swiping to the right,pressing to fold and spread fingers.Then based on the smart wearable device to control the interaction scene of other IoT devices(of course,it is also possible to directly interact with the wearable device itself),we designed a sys-tem that recognizing human arm gestures using the inertial sensor(IMU)in the smart wearable device.By exploiting the sensors in the smart wearable device to perceive the attitude of the wrist,the trajectory of the arm movement is predicted.And the tracked arm movement trajectory is used to classify and recognize gestures,such as raising the wrist,waving up and down,and waving left and right.Compared with sound perception,the replacement of system application scenarios perceived by IMU sensors can provide higher recognition accuracy.