Gait Analysis And Balance Research Of Lower Limb Exoskeleton Based On Multi-Source Information

The lower limb exoskeleton robot is composed of human and external mechanical structure.It involves many disciplines,such as biomechanics,robotics,information science,and artificial intelligence,it can provide additional energy to assist the body to perform actions that can't be accomplished only by natural forces.In recent years,the research of exoskeleton robot has been increasing day by day.The exoskeleton robot can provide the auxiliary force and support force for the wearer to a certain extent.The exoskeleton robot has been applied in many fields.The control signal of the exoskeleton robot has been the focus of research.The surface electromyogram(sEMG)can obtain the movement intention of the human body more easily,the lower limb exoskeleton based on sEMG has an important application value in the field of rehabilitation medicine,it can be applied to the lower limb exoskeleton,which can play an important role in the control signal.In this thesis,the sEMG signal,the plantar pressure signal and the inertial signal are introduced into the control of the lower extremity exoskeleton,and this study includes:gait pattern classification based on multi-source information of lower limbs,establishment of gait model based on PSO-ELM and analysis of lower limb balance strategy and other issues.In this thesis,the sEMG signals of the lower limbs,plantar pressure signals and inertial signals are collected,then the three signals are filtered and extracted respectively.Then,the appropriate kernel function is selected for each signal feature.The optimal combination of kernels is obtained,and multi-core correlation vector machine(MKRVM)is used to classify lower limb gait patterns.Although the training time and recognition time are slightly higher than single-core RVM,MKRVM integrates the features of multi-source information and has higher recognition rate.When the lower limb exoskeleton has recognized different gait patterns,it is necessary to adjust the joint angles,angular velocities to satisfy the current state of motion.In this thesis,a gait model is established based on PSO-ELM,and the parameters of the sEMG signals are used to predict the knee and hip joint angles,the knee and hip joint angular velocities,and the two models are contrasted.In the study of the lower limb exoskeleton,the study of joint torque in balance control plays a key role in the stable walking and balance recovery of the exoskeleton.In this thesis,the sEMG signal is introduced into the analysis of lower limb balance strategy.And the models are simplified for the ankle balance and step equilibrium,and the corresponding joint torques are obtained.Finally,the sEMG and angle signals are used to predict the joint torques,and the theoretical foundation is established for the balance control of sEMG.