Human Motion Monitoring And Rehabilitation Evaluation System Based On Flexible Strain Sensors

Flexible strain sensors are a research hotspot in the field of wearable electronic devices due to their high flexibility,lightweight,and low power consumption.A large number of studies have been reported on developing accurate,fast,and stable strain sensors for human motion detection and physiological signal monitoring.With the rise of Internet of Things and artificial intelligence technology,the design of wearable systems based on flexible strain sensors has rapidly developed.In this paper,a flexible strain sensor is fabricated for motion monitoring and rehabilitation evaluation.A step frequency monitoring and Brunnstrom state judging system are designed and implemented based on this sensor.Our main contributions are as follows:A waterproof CNT flexible strain sensor using latex as the substrate and encapsulation material and CNT as the sensing material is designed and prepared.Through SEM and EDS characterization analysis,it is proved that the device forms a structure with a flexible substrate and package on the outer layer and a sensitive film on the inner wall.The optimal concentration of CNT is determined to be 10 wt% and the optimal time of injection immersion to be 15 minutes through strain performance tests.The device exhibits good linearity in the strain range of 0.1-65% and 65-200%,with response and recovery times of 250 ms and 333 ms,respectively.The sensor has a sensitivity of 21.09 in the strain range of 0.1-65% and a sensitivity of 91.15 in 65-200%.The device has excellent stability and repeatability under large strains,maintaining a stable response during 4000 tensile and release cycles.A step frequency monitoring system is designed to read,analyze,and display the step frequency signal.When the tester equipped with the sensor moves,the upper computer displays the sensor resistance curve,real-time step rate,motion state,and total step frequency count.A rehabilitation evaluation system integrating flexible strain sensing array and Temporal Convolutional Network(TCN)is designed,and a Brunnstrom stage evaluating system that can collect data through portable flexible wearable devices and submit the data to a deep learning model for rehabilitation rating judgment is proposed.A Brunnstrom staging classification model based on TCN is constructed,the signal curve is signal-aligned by low-pass filter and dynamic time warp processing,and the array data signal is extracted by the TCN layer,and then the features are classified by the Softmax layer.The results show that the Brunnstrom staging classification in the test set can reach97.52%,and the accuracy of the model for each grade can exceed 95%,and the test results prove that the Brunnstrom staging judging system based on flexible strain sensor array and time convolutional network can judge the rehabilitation rating of the tester’s hand motor task data.