Measurement And Application Of Knee Joint Angle Based On Flexible And Inertial Sensors

The knee joint plays a very important role in the body’s stable support and walking movement.Its flexion/extension movement can reflect kinematic information such as lower limb muscle strength and center of gravity balance,which is of great significance for medical evaluation and rehabilitation.Now,human motion capture systems based on optical high-speed camera equipment can achieve high-precision measurement of knee joint angles.However,due to its high cost and cumbersome experimental process,it cannot be widely used.Since the21 st century,with the rapid development of electronic technology,sensors have gradually become lightweight,miniaturized and intelligent,and the research of limb capture system based on wearable devices has become a hot topic.While maintaining the advantages of portability,comfort and low price,improving the accuracy of the wearable measurement system as much as possible is a problem worthy of study.Under the above background,this paper establishes the electro-mechanical viscoelastic model of the flexible sensor;on this basis,it uses Kalman filter to fuse the signals of the flexible and inertial sensors to obtain the knee joint angle;and uses the angle discrete sequence as the input for abnormal gait recognition.The specific work of this paper is as follows:Aiming at the problem that the empirical mechanical cannot explain the resistance relaxation phenomenon of the flexible sensor from the mechanism,this study gives the linear spring element in the Zener model with electrical properties to establish an electro-mechanical viscoelastic model.The experimental results show that the model can quantitatively describe the resistance relaxation curves under different strains,and explain the phenomenon that the coefficients in the expression of resistance relaxation are approximately linear with strain.Finally,according to the model,the resistance relaxation curves under low strain condition are used to predict the resistance relaxation curve under high strain condition,which further proves the correctness of the model.In order to solve the problem of hysteresis error in traditional measurement of knee joint angle with flexible sensor,the elastic model that characterizes the conversion relationship between resistance and strain is replaced with the Zener electro-mechanical viscoelastic model.Combining with the information of the flexible and inertial sensors,the Kalman filter algorithm is used to estimate the parameters of each component in the model in real time,and the form variable of Voigt body is calculated to compensate for the hysteresis,and the drift error of inertial sensor is reduced in the process of angle fusion.The knee flexion experiments show that the root mean square error of the angle fusion scheme based on Zener viscoelastic model is better than single sensor measurement solution.Taking the discrete sequence of knee angle as input,an abnormal gait recognition scheme based on support vector machine(SVM)is designed.Firstly,the dual threshold plus time window method is used to recognize the gait period of the input data,and the support phase and swing phase are divided.Then,according to the characteristics of trembling gait,circle gait and scissor foot gait,time-domain feature parameters are extracted from the support phase and the swing phase.Finally,SVM is used to complete gait classification.The gait experiment results show that the recognition rate of the multi-sensor angle fusion scheme is higher than that of the angle measurement scheme based on the flexible or inertial sensors respectively,and indirectly proves the superiority of the angle fusion scheme.