Research On Human Intention Recognition And Motion Control For Lower Extremity Exoskeleton Robots

Wearable powered exoskeleton robots integrate mechatronics,biomechanics,motion gait planning,robotic control and human body sensor network,etc.It is a kind of high-tech intelligent product,and its application demand in medical,military and industrial fields is increasing gradually.In order to improve the performance of lower extremity exoskeleton robots and make them provide more effective help for wearers,this paper studies the trajectory learning algorithm based on dynamic motion primitive and reinforcement learning,and applies the human-computer interaction technology which is based on bio-electricity signal(electroencephalogram and surface electromyography)to the control of lower extremity exoskeleton robots.The main research work of this paper includes the following three aspects:(1)Design of the experimental platform: According to the requirements of application background,this paper designed a lower extremity exoskeleton robot,which considered the research needs and costs to select the appropriate equipment,and combined engineering,mechanics and somatology to design the mechanical structure and some necessary safety measures.Meanwhile,a simple interface was also developed based on MFC to help the operators to conduct experiments and observe results.(2)The movement control of lower extremity exoskeleton robots based on dynamic motion primitive and reinforce learning: In this motion control strategy,to keep the robot in a stable state,the inverted pendulum model was used to build a motion model for the lower extremity exoskeleton.Then by utilizing this model,a motion trajectory was generated which was learned and modeled by the dynamic motion primitive algorithm.At the same time,the reinforce learning based on path integral study was adopted to optimize the trajectory in order to eliminate the influence of the uncertainty in joint space.Experimental results show that this motion control strategy can effectively suppress interference and eliminate uncertainty.(3)The lower extremity exoskeleton control based on bio-electricity signals: In this paper,the human-computer interaction technology based on EEG and s EMG was applied to the control of exoskeleton robots.Firstly,the motion imagination EEG of subjects was decoded as the control instructions that could be recognized directly by the exoskeleton.Then,the intensity value of s EMG collected on the wearer's upper limb was extracted,and used to adjust a trajectory that was generated and constrained by three key points.The two contributions of the strategy of controlling the exoskeleton using the mixed signals proposed in this paper are to free the hands and feet of wearers,and also,enable the exoskeleton to adapt to different types of stairs due to the real-time adjustment of the gait trajectory.The experimental results indicate the control strategy of the lower extremity exoskeleton based on bio-electricity signals makes the exoskeleton more convenient and effective when helping humans.