System Design And Gait Analysis Of Exoskeleton Robot Assisted In Lower Limb Rehabilitation

Spinal cord injury is a serious disabling disease,which often leads to dysfunction such as limb perception and movement below the injury segment.Severe people are more likely to be completely paralyzed.Rehabilitation training can help patients restore physical functions such as movement and perception,improve and enhance patients’ quality of life.However,the current number of rehabilitation therapists in China is limited,far from meeting the needs of many patients.In order to solve the problem that patients with spinal cord injury cannot lock the knee joint autonomously during the gait rehabilitation training stage,and need rehabilitation therapists to assist in locking their knees,commissioned by Kunming Tongren Hospital,this paper designs a lower limb rehabilitation training exoskeleton robot to assist patients with spinal cord injury to restore lower limb muscle strength and the ability to walk in a natural gait posture,the design of software and hardware systems and the research of gait phase recognition methods were carried out.The main research contents are as follows:1)According to the actual needs of gait rehabilitation training for patients with spinal cord injury,the overall scheme of assisting the exoskeleton robot system of lower limb rehabilitation training was determined.In order to reduce the energy consumption during the patient’s movement,the drive system is placed on the mobile support platform,and the motor driving force is transmitted over a long distance by using the Bowden rope;In order to solve the problem of uncontrollable and unstable plantar pressure in patients with spinal cord injury,a plantar sensing module is designed to measure the contact status information between the foot and the ground.Based on the design objectives of exoskeleton robots and the research results of the human lower limb movement mechanism,the selection of motors,sensors and controllers,as well as the design of mechanical structure,sensing system and embedded control system were carried out.2)Based on the sensing system of exoskeleton robot,a gait phase recognition method that integrates the information of the plantar touch state and the information of knee angle and angular velocity is proposed.Firstly,combined with the gait rehabilitation training of spinal cord injury patients and the RLA(Rancho Los Amigos)gait analysis method,the gait cycle was divided into 4 phases,and the human gait data in various exercise modes were collected.Then,K Nearest Neighbors(KNN),Decision Tree(DT)and Support Vector Machine(SVM)model based on Particle Swarm Optimization(PSO)optimization were used to carry out multiple sets of gait phase recognition experiments between different velocities and different individuals,and fault tolerance performance verification experiments.The results show that the proposed method can effectively identify gait phase,the average recognition rate between individuals at different speeds and between individuals is 99.824%.And the experimental results of fault tolerance performance verification show that when the plantar perception module is not triggered,the recognition accuracy of the locked knee phase and the loose knee phase can still reach 94.488% and 91.853%,respectively,which verifies the fault tolerance of the proposed method.3)In order to provide accurate knee lock and knee loosening time points for exoskeleton robots,a collaborative control strategy based on decision tree is proposed to integrate gait phase recognition into system control.A prototype of an exoskeleton robot assisted in lower limb rehabilitation training was built to carry out online experimental research on knee locking/loosening knee at an asynchronous speed.The results showed that the exoskeleton system was able to stably complete the movements of locking and loosening the knees.