Wearable Elbow Joint Angle Measurement System And Research On Accuracy Optimization

The elbow joint is one of the most critical joints in the human body as it is responsible for the forces and stresses between the upper arm and forearm.The flexion/extension angle of the elbow joint and the internal/external rotation angle of the forearm are essential indicators to assess the motor function of the elbow joint.Measuring the elbow joint angle has significant research significance and application value in clinical applications,sports medicine research,ergonomic research,sports training,and athletic performance evaluation.Compared to currently-used vision-based angle measurement system,wearable angle measurement system offers advantages such as ease of use,comfort,making it a more suitable solution for daily life.As ingredient sensors for wearable angle measurement system,integrate inertial sensors and flexible strain sensors have been frequently implemented.However,both sensors face limitations in accuracy due to sensing mechanisms.For instance,flexible sensors are observed with resistance relaxation during measurement due to the viscoelastic properties of the sensing material,resulting in hysteresis errors.On the other hand,inertial sensors receive accumulate drift errors over time,leading to reduced measurement accuracy.To address the above problems,this thesis establishes a force-electro viscoelastic model to describe the resistive relaxation characteristics of flexible sensors.Extended Kalman filter algorithm was then utilized along with the model to fuse the measured angles of flexible and inertial sensors to provide the revised elbow joint angle.These work can be detailed in 3 parts:1)A wearable elbow joint angle measurement system,which is easy to wear,lightweight and comfortable,was built.Sensor selection and installation position determination were introduced,along with data acquisition system hardware design,lower computer program composition,and flexible wearable system construction.2)A second-order Zener parallel force-electrical viscoelastic model that can effectively describe its resistive relaxation characteristics is developed for the fabricflexible strain sensor.Tensile and relaxation experiments were designed,and the experimental data were used to identify the model parameters.The amount of resistive relaxation is represented by the amount of change in the length of the effective conducting region due to the change in displacement of the viscous pot unit in the model.The results show that the predicted resistance change of the identified model is in good agreement with the measured resistance,with a coefficient of determination above 0.98 and a root mean square error less than 0.55 kilohm,which can better describe the viscoelastic characteristics of this sensor.Finally,an extended Kalman filter algorithm is established based on this model to perform resistance compensation by predicting the displacement change of the viscous pot element under different strain states;and the strain prediction results based on the compensation resistance are compared with the actual strain,and the prediction accuracy is above 0.98,which indicates that the established model can be used for strain measurement based on resistance data.3)In order to compensate the relaxation of the resistance of the flexible sensor in the elbow flexion angle measurement process in real time and reduce its hysteresis error,the viscoelastic model and extended Kalman filter algorithm are used to estimate the displacement change of the viscous pot unit in real time with the flexible and inertial sensor information,and then obtain the fused elbow flexion angle to reduce the drift error of the inertial sensor.After the experimental verification of elbow joint angle acquisition,the angle fusion scheme reduces the hysteresis error of the flexible sensor measurement scheme;and reduces the drift error of the inertial sensor measurement scheme in the longtime measurement service.