Research On Flexible Wearable Decoupling Measurement Technology For Human Wrist Joints

Wrist joints are more complex in human joints and also play a very important role in human upper limb movements.At present,many scenes such as sports and fitness,medical and health,and virtual reality entertainment have a huge demand for accurate measurement of wrist motion.However,most existing anthropometric systems have the disadvantage of being uncomfortable to wear and not portable.In recent years,the flexible wear system of human joints integrated with flexible silicone rubber sensors has received more and more attention.This flexible wearing system has the advantages of being comfortable and light to wear,free from site restrictions and environmental interference,and easy to carry,so it has broad development prospects.However,most of the current flexible wearable systems can only detect single-degree-of-freedom movements of human joints.In fact,for a multidegree-of-freedom wearable measurement system such as a wrist,the movement of each joint of the joint causes a strong coupling between the modules of the measurement system,but there is currently no research on this coupling problem.Facing this engineering requirement,this article studies the corresponding coupling law and decoupling technology.The main contents are as follows:In order to minimize system coupling,a low-coupling design for the wearable measurement system.When designing the structure of the wrist flexible wearable system,the actual coupling degree of the system can be reduced through theoretical calculation,and the development difficulty of the subsequent decoupling algorithm can be reduced.Therefore,this paper proposes a method of optimizing the sensor placement based on the analysis of the deformation field of the clothing carrier of the flexible wear system.The deformation distribution of the clothing carrier during joint motion is calculated to determine the coupling degree of the system.In order to obtain the deformation field of the clothing carrier,this paper builds a corresponding visual acquisition system,which makes the clothing carrier have speckle lines through water transfer technology.The deformation law of the clothing carrier is calculated by digital image correlation method,and the sensor is arranged according to this law Carry out optimization,and the optimized layout plan makes the system have a smaller degree of coupling.Research on the calibration of the wearable measurement system is carried out,because accurate wrist calibration measurement system calibration is the key link in constructing the coupling model of the wrist flexible wearable system.This paper designs the corresponding flexible sensor acquisition circuit and communication method.According to the calibration requirements,the hardware of the vision system was actually selected,a vision measurement system was built,the open CV vision library was used for image processing,and the human joints were spatially tracked according to the spatial positioning algorithm.According to the results of spatial positioning image processing,the motion angle of the joint is solved.And compared with the commercial dynamic capture system,it proves that the experimental measurement system built is usable.The coupling problem and decoupling algorithm of the wearable measurement system are studied.Model the joint motion of the wrist joint.By comparing various models,the LSTM model is selected as the coupling model of the wrist wearable system.The corresponding LSTM model was built,some of the hyperparameters of the model were determined according to the experiment,and the learning method of the model and the method of preventing overfitting were analyzed.Collect the data and train to get the LSTM coupling model of the wrist multi-sensor.According to this model,decouple the flexible wearable system.The decoupling results are compared with the real joint motion,which proves the effectiveness of the model.According to the engineering requirements,the self-correction method of the model after the relative position of the flexible clothing and the human body is changed is carried out.Based on the Fine-tuning transfer learning method,the original calibration model can be quickly adapted to the current state,and the model is automatically corrected.