Design Of Wearable Plantar Pressure Acquisition Device And Research On Abnormal Gait Recognition Method

Walking is one of the most basic forms of human movement,and plantar pressure has an important impact on the stability and balance of walking.Uneven distribution of plantar pressure may lead to abnormal gait,excessive plantar muscle tension or fatigue,plantar pain and other problems.These problems can seriously affect people’s quality of life.Plantar pressure collection devices can monitor plantar pressure distribution,identify abnormal gait,assess foot health,and prevent foot injuries and diseases,and have a wide application prospect and demand.However,existing devices have deficiencies in precision and accuracy,data analysis and processing,which need more in-depth research and exploration.To address the above deficiencies,this paper designs and implements a high-precision plantar pressure acquisition device,and conducts preliminary applications and tests on it.The specific research contents are as follows:1.In this paper,a wearable plantar pressure acquisition device based on a flexible thin-film pressure sensor is designed for high-precision plantar pressure data acquisition.The overall structure of the device was designed based on the human foot anatomy and footprint method.Meanwhile,the paper designed and wrote the driver of the plantar pressure acquisition device to realize the function of data export from the sensor to TF card.2.In order to verify the reliability and accuracy of the sensor,the paper designs an experimental device for calibration and verification experiments.The calibration includes static calibration of planar steel surface and static calibration of planar flexible surface for adjusting the voltage versus force relationship of the thin-film pressure sensor on the flexible surface.The validation experiments include dynamic validation experiments on the planar rigid surface,static validation experiments on the curved flexible surface,and dynamic validation experiments on the curved flexible surface.The experimental results show that the thin-film pressure sensor can work well on the footprint with high accuracy and stability.3.In order to verify the reliability and validity of the collected plantar pressure data,the plantar pressure data were visualized by using the cloud pattern map.Firstly,AD data decoding and Kalman filtering were performed on the plantar pressure data,and then the plantar pressure cloud pattern map program was written using MATLAB software.The analysis and comparison of the plantar pressure cloud pattern map for walking on flat ground showed that the plantar pressure cloud pattern characteristics and the gait cycle characteristics could correspond,thus verifying the reliability of the collected data.4.In this paper,a deep neural network is designed and built for abnormal gait recognition.A CNN-LSTM algorithm combining convolutional neural network and long and short term memory network is used to classify the plantar pressure data of six gait patterns.The algorithm combines the spatial feature extraction ability of convolutional neural network and the modeling ability of long-and short-term memory network for temporal data,and can effectively identify abnormal gait with an accuracy of 93.68%.