Research On Motion Monitoring And Rehabilitation Evaluation Based On An Inertial Wearable System

With the development of inertial sensing technology,inertial wearable sensing systems are widely used in medical rehabilitation,sports,fall protection,games,and entertainment.For example,clinically,patient motion biomechanical parameters are quantitatively analyzed by inertial sensing systems to accurately measure and evaluate patient recovery progress and health function;in daily fitness training,inertial wearable sensing systems are used to accurately capture limb movements,output joint motion parameters and displacement of each limb joint segment to quantify changes in athletes’ limb activities and qualitatively analyze them to objectively evaluate training performance and improve training quality.Although there are many studies based on inertial wearable sensing systems,there are still some problems,for example,in fall warning,the system sensitivity is low,it is difficult to distinguish between real and fake fall movements,and the lead time of fall warning is short,and the accuracy rate is not high;in rehabilitation motion monitoring,the clinical nature is not enough,the monitoring data fluctuates a lot,and the accuracy of motion capture is low;in rehabilitation assessment,the assessment scheme is single and the target is weak.To address the above problems,this paper conducts a study on the inertial wearable system-based rehabilitation motion monitoring,evaluation,and fall warning for clinical patients after total knee arthroplasty(TKA).First,a multifunctional wearable inertial sensing system based on a high-performance multi-protocol SOC communication chip n RF52832 was constructed.The design uses the MPU9250 nine-axis inertial sensor motion acquisition module,adaptive multi-inertial sensing data transmission data frame encapsulation protocol,communication,voice hardware,and software modules.The whole system adopts the architecture design of central micro host-multimotion measurement nodes,and the host and nodes communicate with each other by wireless Bluetooth self-organizing network,forming a low-cost,convenient,and monitoring flexible high-performance inertial wearable sensor network system.To address the problems of high false alarm rate,low accuracy,and short lead time of fall warning in fall warning research.By characterizing the fall action and daily fall-prone hazard action,this paper designs a fall warning algorithm based on offset displacement,which finely divides the fall action many times by calculating the fall weightless displacement offset.Firstly,the backward,left,right and daily fall-hazardous actions are distinguished by the combined acceleration and posture angle respectively,and the human motion is divided into daily and falllike actions.Then,the human body weightless bias displacement is used to further divide the fall-like motions to obtain the true fall and false fall motions.Experiments show that the system has 98.6% recognition accuracy and a low false alarm rate of 1.4% for daily fall-prone dangerous movements,98% average fall warning accuracy,96.1% sensitivity,strong fall motion detection capability with 99.2% specificity,and excellent performance in differentiating daily fall-prone dangerous movements from fall movements.The overall complexity of the algorithm is low,which is suitable for system embedded edge terminal operation and provides excellent performance in fall motion detection and warning for the whole machine.To address the problems of large errors in rehabilitation motion monitoring and the lack of targeted and effective rehabilitation assessment programs.Firstly,the data fusion algorithm based on complementary filtering was improved,and the combined acceleration function was innovatively introduced as the weighting coefficient to measure the accelerometer and gyroscope errors to improve the accuracy of quaternion calculation,and to play a good role in curbing the interference of strong forces over a longer period and eliminating the interference of motion artifacts.Then,the joint rehabilitation activity estimation algorithm for TKA patients is designed,and the method of double matrix internal rotation superposition is used to estimate the joint activity angle,which avoids the error caused by the posture angle process volume and improves the joint angle estimation accuracy compared with the direct operation of the posture angle volume.Finally,the clinical treatment information is tapped,and the gait flexion and extension of joint mobility of TKA patients are used as indicators to assess the degree of joint rehabilitation of patients to enhance the accuracy of joint rehabilitation assessment.The experiment showed that the mean error of the monitoring results of this system and the BTE device(authoritative equipment for rehabilitation assessment)on the static joint flexion and extension activities of TKA patients was 4.17°,and the mean error of the results of the healthy group was 0.52°,and the absolute difference of all subjects was between-10° and 10°,which met the error criteria for clinical use.Meanwhile,the degree of rehabilitation of 10 patients was comprehensively analyzed in terms of both active/passive joint flexion/extension and gait flexion/extension activities,and the assessment results were consistent with the actual clinical performance.The whole machine showed excellent clinical performance in the rehabilitation exercise training and rehabilitation degree assessment of TKA patients after surgery.