Research On Data Glove For Hand Rehabilitation

Traditional hand rehabilitation has many problems such as high price of rehabilitation equipments and lack of medical resources.In order to assist doctors to evaluate hand rehabilitation process,make reasonable hand training plan and guide patients with long-term rehabilitation at home,this thesis puts forward a kind of data glove system for hand rehabilitation,including the design of data glove system,calibration of sensors on data glove,data fusion and attitude calculation,using data glove to conduct human-computer interaction and medical index judgments of hand.The main contents and achievements are listed as follows:(1)In order to collect hand information with low-cost and high-accuracy method,data glove system with MPU9150 inertial sensors is developed.According to system requirements,the hardware design is realized to collect acceleration,angular velocity,and geomagnetic information from main joints and the back of hand.(2)In order to get attitude angles of main joints and the back of hand,data from sensors are calibrated,fused and calculated.The zero deviation of gyroscope is estimated by six position method.The calibration parameters of acceleration are calculated by the least square method.The hard magnetic interference and soft magnetic disturbance of magnetometer are reduced by ellipsoid fitting method.Data fusion of acceleration,angular velocity and geomagnetic data is carried out by complementary filtering based on quaternion.Finally,the attitude is calculated by quaternion.The experiment shows that the static accuracy of attitude can reach 0.3 °,dynamic accuracy can reach 1.7 °.(3)In order to conduct real-time and high-accuracy human-computer interaction,virtual hand model and virtual scene is established in Unity platform,so the glove data can map to virtual hand joints and the attitude can convert to rotation.The script is designed to control the hand model in Unity.The experiment shows that the data glove in this thesis can control the hand model in Unity to conduct real-time and high-accuracy human-computer interaction.(4)In order to offer doctors a reference for hand evaluation,the hand evaluation of medical hand rehabilitation is discussed in this thesis.According to the evaluation of medical hand rehabilitation,range of joints angle,degree of tremor and flexibility of hand are expressed to by angle,stability and operation time.The angle range of hand joints is evaluated by two methods.One is TAM rating method based on medical evaluation standard.The other is gesture recognition method based on BP neural network.The degree of tremor is evaluated by BP neural network,and the hand flexibility is evaluated by action time.The experiment shows that the data glove system can be used to accurately evaluate the range of joints angle,degree of tremor and flexibility of hand.