A Fabric Sensing System For Information Collection And Pattern Recognition Of Lower Limb Movement

The information of human movement acquisition and pattern recognition are the research bases of the judgment of human intention,medical health,VR virtual manmachine interaction,wearable exoskeleton robot control and other fields.In wearable exoskeletons,monitoring and recognizing human movement patterns are particularly important.The premise of wearable exoskeleton motion control is to be able to judge the current motion before proceeding to the next step.Wearable devices are also frequently designed and widely used in healthcare applications.Features of Parkinson's disease can be assessed by analyzing changes in gait,and gait monitoring systems can provide a real-time picture of the patient's status,which can help doctors make an accurate recovery plan.However,the existing gait monitoring system is hard to meet the daily monitoring,and the traditional sensor is fixed at the human body by binding or by adhesive tape.This creates obstacles and discomfort for the body to move.The subject may need a hand-held power supply,acquisition card.These acquisition systems are not convenient.On the other hand,the present study is faced with problems such as single acquisition signal source,small number of subjects,single movement pattern of subjects,and low recognition rate of movement pattern.In order to solve the above problems,a portable clothing system is designed and produced in this project.The sensor network contains multiple signal sources,and nine common lower limb movements are performed on ten subjects.The correct rate of movement pattern recognition is up to 99.96%.The main work of this paper includes the following points:1.This paper designs reasonable sensor network,selects the sensor type and installation location,and collects the motion information of human lower limbs,including plantar pressure signal,angle information of knee joint,hip joint and leg Angle information,etc.,so as to use as few sensor parts as possible and reflect the motion characteristics of lower limbs more comprehensively.2.In this paper,a wearable collection system for lower limb wearable gait monitoring is designed and manufactured.The sensor network is organically combined with clothing to improve the portability of the whole collection system.The first order inertial filter and Kalman fusion method are used in data processing to improve the accuracy of measurement.LabVIEW human-computer interaction interface is designed and made,which can realize real-time monitoring of human lower limb gait information.3.The experiment use the portable collection system to carry out the data collection of human lower limb movement,and collected nine common lower limb movements of 10 subjects.The moving window average value and standardization are used to extract data features,and the classification models trained by different sensor data combinations are compared.The optimal combination of sensor information is obtained from the analysis of multiple evaluation indexes,and the rationality of sensor network design is verified.The optimal information combination and classifier have a pattern recognition accuracy of 99.96% for the nine lower limb movement patterns.