Semg Control The Lower Limbs Exoskeletons Rehabilitation Robots Of Research

At present, in our country, the research of rehabilitation robot for patients with stroke hemiplegia is still in its infancy now. Because of the function obstacle of limbs locomotion, which is caused by central nervous system injury, results in the deterioration of the quality of life, it becomes imperative to design neuro-rehabilitation robot, which will be used to assist patient to recover the locomotion function of affected limbs. Studies show that locomotion training of affected limbs, which is based on active consciousness control, contribute to recover the locomotion function of affected limbs. In this paper, through myoelectricapparatus, sEMG, which can reflect locomotion condition of corresponding muscle, will be used to control the robot and reflect the locomotion intention of patient. The major researches in this paper are listed as follows:First, the etiopathogenisis of cerebral apoplexy and rehabilitation treatment methods are elaborated in detail, the development and application of sEMG are reviewed, and puts forward methods and thoughts of the research. The generation principles of sEMG and movement features of lower limbs muscle are introduced.Secondly, preferred method is raised to choose the useful sEMG, and then the data of sEMG will be analysis, including filtering, linearization, action segment detection, time-frequency domain eigenvalues of the surface EMG are extraction, such as mean value(MAV), wave length (WL), variance (VAR), the number of zero crossing point(ZC), slope change several (SSC), Willison amplitude (WAMP), mean frequency, median frequency, Ratio parameter and so on. After that the dimension reduction and evaluation of those eigenvalues will be done, radial basis function (RBF) neural network and learning vector quantization (LVQ) neural network are used to indentify the movement pattern. Lower limbs rehabilitation robot model, which is created by Solidworks software, will be imported to Matlab and create a file, which can be simulated by SIMULINK, next, VR module of Matlab and SIMULIK module are used to simulation, and then analyze the result. Experiment results show that learning vector quantization (LVQ) neural network is perfect.Finally, two kinds of equipment are used to collect sEMG of lower limb muscle and compared. Experimental results show that sEMG, which has been preprocessed and classified, can driven simulation models, and reflects the movement intention of patients, laid a solid foundation for the design for active conscious control system of lower limb rehabilitation robot.