Gait Acquisition System Based On The Combination Of Wearable Devices And Azure Kinect Sensor

Gait analysis is of great significance in the fields of rehabilitation medicine,identity recognition and sports training.The traditional wearable gait acquisition system will cause large errors in gait parameters due to different sensor wearing positions.The equipment of gait acquisition system based on marking method is expensive and needs to be used under the guidance of rehabilitation doctors in combination with force-measuring system,which is inconvenient for clinical application.At present,the field of rehabilitation medicine lacks a convenient,efficient and inexpensive gait acquisition and analysis device.Based on the existing problems of current equipment and the demand for new equipment,this paper designs a low-cost gait acquisition system.The research work is as follows:In response to the expensive cost of precise labeling devices,this article designs a gait acquisition system that combines Azure Kinect with a single wearable sensor(foot pressure acquisition system),which extracts gait features and joint angle features through the fusion of two sensors.Among them,the plantar pressure information during human walking is extracted through thin film pressure sensors,and the three-dimensional coordinate data of32 joint points during human walking is obtained through Azure Kinect.The hardware design of the foot pressure collection system mainly includes modules such as MCU selection and wireless communication.The upper computer program is designed for wearable sensor devices and Azure Kinect,and the pressure data is transmitted wirelessly,while the coordinate data is generally wired.Analyze the human walking process,complete the human gait modeling and joint simplified modeling,propose the key feature parameters and feature parameter extraction method during the walking process,analyze the plantar pressure changes during the walking process,complete the gait segmentation,process the original pressure data to extract the gait time parameters,and combine the original coordinate data with the pressure data to extract the gait space features and joint angle features.In order to verify the accuracy of the gait acquisition system,a high-precision camera system is designed for gait test,and the error analysis of the feature parameters extracted by the camera system and the gait acquisition system is carried out.The consistency test of the two systems shows that the parameters obtained by the two systems have good consistency(Pearson correlation coefficient r>0.8,P < 0.05)and small error(root mean square error of gait parameters is less than 0.1,and root mean square error of joint angle is less than 6).The upper and lower limb weight-bearing test was designed.Through error analysis and symmetry analysis of the characteristic parameters extracted from the test,the weight-bearing state and weight-bearing state were distinguished.A classifier is designed for the extracted feature parameters.After the feature is input into the classifier,the recognition can be completed with an accuracy of 80%.This confirms the practical applicability of the gait acquisition system studied in this paper.