Prediction Of Posture And Movement State Of Human Lower Limbs

Exoskeleton robots are an important means of extending and enhancing the body's athletic ability.However,the balance of the exoskeleton robot and the human-computer interaction with the human body,especially how to achieve the coordination between the exoskeleton and the body's own movement,is one of the key issues of the planning and control of exoskeleton robots.In this study,we aimed on the method of predicting the posture of lower limb and the method of motion state prediction.The main contents of this paper include:First,based on the fluctuation of the center of the plantar pressure center,the balance control strategy of the human body under static standing is analyzed statistically through designing three experimental variables of the slope,the opening eyes and the time.This paper reveals the partial rule of the pressure center under the change of different variables.Moreover,the single factor analysis of variance is performed on our results to verify the exixtence of significant difference.Second,based on the EMG signal of the muscle of the lower limb muscle group,the EMG-to-motion mapping was trained through the support vector machine.The continuous movement of the lower extremity motor gait and the sagittal plane of the hip,knee and ankle prediction.Through the motion prediction and statistical analysis of ten healthy subjects,it is concluded that the average classification accuracy of the gait class and the continuous movement accuracy of the knee joint under different motion modes are both highest while walking.The experimental results show that it is feasible to predict lower limb motor gait and multi-joint continuous motion based on EMG.Third,in order to verify the practical value of the estimation of the lower extremity joint motion,the subjects were asked to walk on the treadmill for on-line prediction experiments.It is possible to achieve the effect of online real-time prediction by the motion prediction and statistical analysis of four healthy subjects.