Research On Human Posture Recognition System Based On Highly Stretched Strain Sensor

With the advancement and development of science and technology,the detection and processing of human posture information is an indispensable testing technology applied in many fields.In recent years,the rapid development of wearable devices,especially stretchable strain sensors,has provided a completely new solution for human posture recognition.In this paper,rubber polymer with ultra-high tensile properties is used as the stretching substrate,carbon black and single-walled carbon nanotubes are used as conductive fillers.Under the adhesion of silicone rubber,a highly stretched strain sensor for human posture recognition is designed.The effects of the distribution ratio of conductive filler groups and different preparation processes on the tensile properties and electrical properties of flexible tensile sensors were investigated.The microscopic characterization and characterization of a flexible high tensile sensor based on carbon black/single-walled carbon nanotubes/rubber polymer were carried out.The static and dynamic properties of the flexible high tensile sensor were thoroughly studied.The experimental results show that the highly stretched strain sensor based on carbon black/single-walled carbon nanotubes/rubber polymer not only has ultra-high stretch ratio,but also has comprehensive and excellent dynamic characteristics,which is the perfect choice for human body posture recognition applications.Based on the prepared flexible high stretched strain sensor,the signal acquisition and processing circuit is designed.The sensor data is sampled by the sliding time window method,and the characteristic parameters of the human body posture are extracted from the data in the window.BP neural network is used to establish the mapping relationship between human body feature parameters and four different postures.The training samples were selected to train the BP neural network to obtain the human posture recognition model,and the model was tested using the test samples.The experimental results show that the recognition system can achieve a correct recognition rate of 86% for four different human postures.The feasibility of the prepared highly stretched strain sensor based on carbon black/single-walled carbon nanotubes/rubber polymer for posture recognition was verified.