Research On Control Strategy Of Lower Limbs Exoskeletons Rehabilitation Robot

For the problem of Stroke patients with hemiplegia rehabilitation training of the lower limbs,the subject system in-depth studys of the lower limbs of patients with the function and effect ofrehabilitation training, and put forward passive, power, active and resistance rehabilitationtraining mode of lower limbs rehabilitation robot. The paper makes analysis and simulation forgait planning and the control strategy of passive, active and resistance rehabilitation trainingmode. To be specific, they can be described as follows:（1）The human body joint mathe matical models are builded according to the normal hipsand knees motion data. Brunnstromc exercise rehabilitation therapy theory and the changes oflower limb surface electromyography of the training process are combined to plan the humanbody physiology training gait with adaptablity and pertinence for patients. The patient’s ownhealth leg joint movement data is obtained in Visual C++written environment. The complexityof dealing with the data using Mtalab for healer can be avoided, and the fitting result will be thereference model of gait coordination rehabilitation training because of lower limbs with gaitsymmetry. The research of this paper deepens the gait planning method of understanding andcomprehension, and provides reference for rehabilitation training clinical test and searching formore stability reasonable gait planning.（2）To achieve real time and high precision of the control system in the process ofrehabilitation training rehabilitation training process and eliminate the effect of unmodelleddynamics, external disturbance and nonlinear uncertainties in the system, two separatecontrollers which contain normal model based computed torque control and variable structurerobust adaptive compensation control are used together. The compensation control is used torestrain the effect of unmodelled dynamics, external disturbance and nonlinear uncertainties byintroducing a dynamic signal and nonlinear damping terms, meanwhile, the design of adaptivecontrol law is used to update the uncertain parameters, enhances the robustness of the controlsystem and ensures global asymptotical stability. This control strategy is proved to be feasibleand effective by Lyapunov stability theory and the simulation results.（3）For the problem of lower limbs exoskeletons rehabilitation robot SensitivityAmplification Control needing accurate inverse dynamics model, the accurate three-dimensional entity model is established by Solidworks, and jointed Matlab/SimMechanics to establish thedynamic model of exoskeletons to realize simulation analysis. After the simulation theinput/output data are trained to get exoskeletons inverse dynamics dynamic mathematical modelby BP neural network, which can ensure sensitivity amplification controller control accuracy.The sensitivity amplification control can reduce the patients bearing capacity and physicalconsumption, realize patients and exoskeleton common movement at the same time acquisitionto accurate human data. The switching of active training mode and resistance training mode canbe realized by adjusting the sensitivity coefficient.