Research On Knee Joint-assisted Motion Control System

Knee assist control is designed to reduce human knee and ankle injuries and provide rehabilitation training for people with damaged lower limbs.The exoskeleton system is controlled by a human body and mechanically provides external force to enhance human strength.The exoskeleton control system is an embedded control system.Followed by the theoretical framework of machine learning and deep learning becomes more and more consummate,some simple and repetitive tasks have been replaced by artificial intelligence products.More and more artificial intelligence theories are widely used in robot control systems.It’s possible to apply artificial intelligence technology in the exoskeleton control system to reduce the coupling between human and machine,and make system more intelligent and comfortable.This project aims to design a wearable knee joint assisted motion control experimental platform for the continuous distribution of plantar pressure and the gait of the lower limbs of the human body during natural walking.The system is divided the control system into three levels for gradually processes data,abstracts,and generates control parameters from the acquisition of the original data.There are low-level controller,a mid-pole controller,and a high-level controller.The main function of the low-level controller is to measure the pressure and the attitude angle of the knee joint,In order to ensure the data is stable and reliable,the data is also processed in the low-level control.The intermediate controller is responsible for the control of the motor,which is controlled by fuzzy PID.The high-level controller is located on the host computer,and the SVM model implemented by Python is used to realize the division of human natural gait,it will output the result,and then generate the actual control parameters for the motor to realize the control method of the motor following the person walking.The overall system hardware consists of linear motors,rotary motors,ESP8266 WIFI modules,FSR pressure sensors,MPU9250,host computer,STM32F407,and the system is integrated through CANOPEN,USART,IIC,TCP / IP,SPI protocols,in order to ensure the stability of the system,Transplant UC / OS-II as a task management system.At the end of the project,the sole pressure is used as the input of the SVM model to output the current walking state,and the dynamic state of the current state of human walking is used to output the PID parameters of the motor to achieve real-time assistance for human walking.The successful implementation of the control system will reduce the inconvenience of users of the exoskeleton system,and subsequent researchers can obtain stable data on the basis of this system and continue to improve.