Research On Preparation Of Flexible Sensor Glove And Gesture Recognition

The posture of human finger reflects movement function and information of hand.Collecting finger movement information has practical application significance in rehabilitation medicine,sports bionics,human-computer interaction and other fields.For example,finger motion information can be used to develop robotic arms,Virtual Reality and somatic games.At present,to generate a smart glove,most research monitoring hand motions is focused on combining rigid or flexible elements with ordinary gloves.In fact,sensors based on metal or semiconductors have limit extensibility,long-term wearing will lead to discomfort.In contrast,wearable knitted sensors are fabricated based on yarn materials and fabric structure,hence,they are soft and elastic to stick with human body closely and sense the degree of finger joints bending.Meanwhile,discomfort caused by long-term wearing can be reduced,which has important reference significance for applications.During this research,silver-plated nylon was used as conductive yarn,electrical-mechanical performance of various knitted sensors was discussed.In addition,an integrated sensor glove was knitted by a computerized flat knitting machine,and its ability to recognize hand gestures was studied.The specific research content is as follows:First,electrical properties of knitted flexible sensors with different elasticity were discussed to select the best material for conductive area of sensor glove.Using various spandex core-spun yarn as plating yarn,three kinds of elastic conductive fabric were knitted by a computerized flat knitting machine.Through elasticity test method with fixed elongation,the influence of different spandex content,washing and ironing conditions on the elasticity of knitted yarn fabrics were explored,and difference in electrical properties of different elastic weft-knitted conductive fabrics under single stretch and repeated stretch conditions were observed.Second,the optimal size specifications of intarsia conductive area and glove to prepare an integrated glove with five sensor areas were explored.On the computerized flat knitting machine,various conductive fabrics were obtained by changing the number of silver-plated nylon yarns and knitted courses of the conductive area.Strain-resistance test was performed on these samples,then,the electrical performance was observed and sensitivity and linearity were comprehensively compared.In addition,according to the hand size data of the same experimenter,three sizes of gloves were knitted and the conductive area was embedded in the index finger area to explore the influence of the allowance of the glove on the finger sensing performance,and determine the optimal knitting size of the sensing glove as well as fabricate it.Finally,the gesture recognition capability of the fully formed sensing gloves was discussed.First,integrated circuits and Bluetooth modules were used to build a five-channel resistance test system and resistance signals of the five conductive areas of the sensing glove in real time were obtained.Second,MATLAB was used to preprocess the training data collected by sensor gloves,the eigenvalues of each gesture was extracted and decision tree form judgment sentences was written in MATLAB to realize static recognition of digital 1-9gesture signals;third,MATLAB’s BP neural network tool was used to train neural network models combining acquired gesture signals and establishes a GUI interface to realize real-time recognition of digital gestures.Through the above research,it is found that washing and ironing have a relaxing and styling effect on elastic fabrics,therefore,elastic recovery rate of the fabric can be improved.Conductive fabrics with better elasticity have excellent sensing repeatability.In addition,the number of conductive yarns,rows of conductive areas and size of the conductive gloves all have an impact on the sensing performance.The five-finger sensing glove based on the optimal specifications had a higher gesture recognition rate,which verifies that the flexible sensing gloves designed in this subject have achieved recognition of gesture signals.