Research And Application Of Interfacial Pressure Distribution Between Below-knee Prosthetic Socket And Stump

The below-knee prosthesis is one of the most common rehabilitation medical instruments in daily life,and it is also the first choice for below-knee(BK)amputees to recover their functions and integrate into the society.The effective optimization of the prosthetic socket can significantly improve the stability of the BK prosthesis and its adaptability to human stump.The mechanical characteristics of the interface between the prosthetic socket and the stump are the most direct,most credible and most stable interaction information generated in the man-machine interaction process.Therefore,based on the dynamic pressure distribution of the contact interface between the prosthetic socket and the stump and gait characteristics,this paper studied the load distribution mode of the residual limb terminal under different motion modes and the influence of different load distribution on the gait posture,and proposed the optimization scheme of the socket.In addition,the application of residual limb load distribution in different motion pattern recognition and prediction of below-knee prosthesis was explored,and a new gait recognition and prediction method for below-knee prosthesis was proposed.By using the self-designed flexible thin film pressure sensor network,the pressure values of the ten typical bearing areas of the contact interface between the prosthetic socket and the stump were measured in real time and synchronously in five typical motion modes,namely,walking on a flat road,going up and down stairs,and going up and down slopes.At the same time,the posture of the knee joint was measured by using the nine-axis attitude sensor.According to the pressure values of residual limb terminal under different movement modes,the law of interfacial pressure change was analyzed,and the influence of the pressure change at different bearing areas on the posture of the prosthesis was studied.An expression method for representing the load capacity of the residual stump was established by using the idea of integration,and the load distribution of the residual stump was represented by the load amounts at different points of the stump.The load distribution patterns of residual limbs in each gait under different motion modes were analyzed.Finally,a numerical model of motion pattern recognition and prediction based on BP neural network with variable learning rate momentum was established by using the load distribution model of the stump,and good results were obtained through data training.Based on the research and analysis,the main conclusions are as follows:(1)The pressure variation characteristics at the interface between the prosthetic socket and the different bearing position of the stump are different in the process of movement.In the five movement modes,the left side of the tibia and the rear side of the stump were the main bearing areas,while the load on the right side of the tibia was small.The overall load distribution on the stump was asymmetrical,and the asymmetric load distribution of the residual limb was more prominent in the up-slope movement mode.(2)The actual gait posture of the volunteer and the posture data of the knee joint confirmed that a large pressure difference between the bearing area on the two sides of the tibial can cause the prosthesis to tilt.The load bearing capacity of the rear tibial bearing area was significantly higher than that of the front tibial bearing area in the five movement modes,which is in line with the bearing law that the deformation of soft tissue is proportional to the load capacity.In addition,the load bearing capacity of the front tibia and the rear tibia were both larger in the up-down slope movement mode,indicating that the movement performance of the BK prosthesis in the up-down slope needs to be improved.(3)Based on the BP neural network of gradient descent algorithm and the variable learning rate momentum gradient descent algorithm,a numerical model was established to identify and predict motion patterns by using the load distribution of residual stump.The simulation results show that the neural network numerical model based on the variable learning rate momentum gradient algorithm can well identify predict corresponding movement patterns,which confirms the feasibility of the identification and prediction of the prosthesis posture by using the interface pressure distribution between the prosthetic socket and the stump.The results provide a new technical route and a reference method for the control mechanism of below-knee prosthesis and the design and optimization of the prosthetic socket.