Gait Recognition Of Biomimetic Prosthesis Foot Touching Ground Based On Stress-strain Principle

Prosthetic foot is an important part of lower limb prosthesis,and its performance is of great significance to the gait and daily life of lower limb amputee patients.However,the existing commercial prosthetic feet are mostly passive,unable to provide the feedback of touch information to the wearer,and unable to play the sensory feedback similar to human feet.In this paper,the strain sensitivity of the prosthetic foot to the ground was analyzed based on the finite element method.The steel column was applied to different parts of the foot to simulate the gait characteristics of the normal foot during the ground contact,so as to find sensitive areas of the foot stress and strain.Based on the strain testing system of prosthetic foot,the static and motion state tests were carried out to determine the best position and scheme of strain gauge sticking at an angle of 30 degrees with vertical direction.With the help of Vicon optical 3D motion capture system,Kistler force plate and strain measurement system,the touchdown kinematics test and strain test of amputee patients were completed,and the kinematics,mechanical action characteristics and strain information characteristics in the touchdown cycle with asynchronous speed were analyzed: Because of the impact,when the heel contacts the ground,the foot is subjected to gradually increasing tension,and the strain increases linearly to the peak value.When the amputee is in the middle of touching the ground,the back foot first touches the ground,the prosthetic foot is pulled,and the strain gradually decreases to zero.After that,the forefoot touched the ground,the foot was pressed,and the strain increased linearly to the trough value.In the stage of stepping on the ground and toe off the ground,the pressure on the foot decreases gradually because the front part of the foot pushes on the ground and the foot rises suddenly,and the strain decreases linearly to zero at this time.In order to quantitatively and intuitively analyze the relationship between strain and gait characteristics,this paper uses machine learning method to intelligently identify touchdown gait characteristics,through the ground reaction force(especially in vertical direction)to classify the gait feature data,could be divided into three touchdown during gait phase,to extract the data,the slope of strain and the average three characteristic values for gait recognition,the five classification learning algorithm.The results show that the accuracy of multi-layer neural network based on touchdown gait recognition is 95.90%,and the recognition rate is generally high,and the recognition accuracy is stable,which can be used as a machine learning model for gait recognition during touchdown identification.Therefore,the stress-strain principle can effectively identify the gait characteristics of amputee patients during touchdown,which provides a new idea and method support for the intelligent application of bionic prosthesis.