Research On Integrative Wearable Devices Based On Flexible Thin-film Sensors

With the rapid development of big data,artificial intelligence,and Internet of Things technologies,wearable devices have been fully developed and applied in medical health monitoring.Wearable devices for real-time exercise data,health monitoring and human-computer interaction are widely used in our daily life,making it a lifestyle that people are keen on.By introducing flexible sensors into wearable devices,the problems of poor wearing comfort,high preparation cost,and poor flexibility in rigid MEMS sensors commonly used.Consequently,flexible sensors are considered to be a key component of the wearable devices in next generation.However,the fabrication methods of the sensors are complex with low preparation efficiency,and they show poor stability after longterm use.In addition,the types and functions of wearable devices based on flexible sensors are relatively simple.As a result,there a still many challenges facing for the flexible sensors applied in wearable devices.Therefore,in this work,we proposed a flexible thin-film sensor prepared by a facile and scalable method,which exhibits ultra-high sensitivity,excellent stretchability,and considerable stability.Based on this,the patch and fabric wearable devices were further designed and fabricated,by which we realized the monitoring to bending of different joints,and accurate recognition to 19 static gestures.The relevant work and results are as follows:(1)We prepared an MXene/PU composite by aqueous mixing and dispersion,and then a facile and scalable brush coating was adopted repeatedly to produce a composite film on various flexible substrates(such as Natural Rubber).Furthermore,we fabricated an ultrasensitive strain sensor by connecting the composite film on the substrate with conductive carbon electrodes.The gauge factor(GF)of the sensor(@35%strain)has impressively achieved 6354,and it made a response to tiny deformations(@0.5-3%strain),which showed the ultra-high sensitivity of the sensor.(2)Furthermore,based on MXene/PU composite film,we prepared smart patches on rubber substrates and fabric devices on wearable items like a rubber glove,fabric wristband,and elbow pad by the brush coating method.After testing,the GF values of the smart patches to the bending of the wrist,knee,and finger joints at degrees of 45°,60°,and 80° were 26.85,21.04,and 27.97;The GF values of the fabric devices to the bending of the wrist at degrees of-90° and 90° were-41.13 and 127.11 respectively.In addition,we adopted the smart patches to monitor the process of exhalation and inhalation,and the wrist pulse successfully.As a result,the as-prepared wearable devices based on MXene/PU composite films showed the capability of detecting various physiological signals of the human body.(3)We further designed and fabricated a low-cost,simply prepared,and flexible data glove with integrative sensing units by embedding the electrodes and then coating MXene/PU composite films on a rubber glove.The data glove provides accurate feedback and response to the bending of the finger joints.We further used the data glove to realize the data collection,data pre-processing,and extremely accurate recognition to 10 static gestures representing digits from "0" to "9" in American standard,and the average recognition accuracy has impressively achieved 99.13%in 3000 test samples.(4)Finally,we used the as-prepared data glove,fabric wristband,and elbow pad to monitor the motions of the human arm and realized perfect recognition to 19 static gestures(including digits from "0" to "9",and 9 complex gestures primarily differentiated by flexion-extension of wrist and elbow joints),in which the recognition accuracy has impressively achieved 100%in 5700 test samples in total.Furthermore,we designed and developed a graphical user interface(GUI)system for application of the above gesture recognition device,which can recognize the gesture of the user with the fabricated data wearable devices in real-time by applying the above trained MLP algorithm model.The recognition results are presented by voice and image.