The Rehabilitation Training Control System Research Of Lower Limb Rehabilitation Robotic

With the aged society,the number of paralysis of lower limbs in elderly people is increasing,so the robot-assisted rehabilitation and therapy has been developed and researched.In this paper,the control theory of the rehabilitation robotic is researched.It includes:(1)According to analysis of the recovery theory,the existing rehabilitation pattern and the rehabilitation strategy,we proposed a rehabilitation training control system is conducted by the Physiotherapist(PT).The rehabilitation program includes the Gait Pattern Predicted Model(GPPM)and adaptive controller.GPPM is based on the patient's exercise activity,PT select the corresponding Stride Length(SL)and Gait Speed(GS)for GPPM input,the gait pattern is generated.In order to maximally promotes patient involvement and strengthen patient's Muscular Activity,put the gait pattern into the proposed adaptive control system.(2)Based on the rehabilitation control system,the GPPM should be designed for the first.The gait pattern is not only decided by kinematics' characteristics of human physiological mechanism,but also influenced by some other factors such as age,gender,etc.To derive the theory of natural gait pattern is difficultly,because of its generation is relatively complex.So the method of experimental analysis was used in gait pattern.The normal walking joint angle data was captured by the technology of video capturing and image processing.The joint angle waveforms of hip and knee in the sagittal plane for the walk of subject curve fitted by the fifth-order Fast Fourier Transform(FFT)function in our work.The joint angle waveforms and SL and GS mapping relationship is difficult to set up,so the Generalized Regression Neural Network(GRNN)is used to build the GPPM.Its efficiency and accuracy was demonstrated by the comparison experiment.(3)After the GPPM built,in order to maximally promotes patient involvement and strengthen patient's Muscular Activity,we proposed an adaptive controller,the Radial Basis Function(RBF)neural networks is used to learn the lower limb exercise capacity of humanrobot,learn the patient's ability and effort based on the errors,maximally promotes patient involvement.In order to strengthen patient's Muscular Activity,so position tracking and stability of the control system should be guaranteed,sliding mode control is applied in our work.According to the Lyapunov stability theory,the stability of adaptive controller is demonstrated.(4)The adaptive controller is demonstrated by a simulation in simulink.The results show that the controller guarantee high robustness,small trajectory tracking errors and maximally promotes patient involvement.The clinical experiment research should be for the future.