| Exoskeleton robot is a highly mechatronic wearable auxiliary robot with bionic structure.The perceptual control system of the exoskeleton manages the driving of the exoskeleton actuator to follow the human body movement in real time by sensing the movement and recognizing intention of the human body.It can enhances the strength,speed and endurance of the human and greatly improves the wearer's ability in carrying heavy load in different environments.Therefore,it has a good application prospect in the military field,medical rehabilitation field and industrial field.So it is necessary to research the perceptual control system of the exoskeleton to make it follow the human body movement in real time,reduce the interference between man and machine,and achieve the goal of smoothly and efficiently assisting human walking with load.This study focus on the research and design of the sensing and control system of the assisted lower extremity exoskeleton robots.Firstly,to meet the requirement of measuring human-exoskeleton interaction,a distributed sensor network and the corresponding communication system was constructed.The sensing system acquires the body attitude and motion data of the human-exoskeleton system by assisting the sensor network laid out on the mechanical structure of the exoskeleton.The denoising and smoothing of the sensor data was pre-processed by using different filters.Then the derivation of the time series prediction algorithm required by the system is completed.By using the correlation coefficient test and the least square method,the model parameters are properly estimated.Also the prediction effects of different order models are verified,and the actual model order is determined.Thereby,the time series prediction of the joint angle of the exoskeleton system motion is completed and the influence of the system delay on the control system is suppressed.Then the identification method based on the foot pressure detection to determine the support state of the lower limbs of the human body is studied.Based on the study of the dynamic equations of the exoskeleton mechanical legs in both the free state and the support state,the study establishes the control method to calculate the moment compensation of the joint driving torque from the exoskeleton joint angle feedback.At the same time,human-exoskeleton interaction force is introduced into the control loop as the feedback.Combined with the predicted angle of exoskeleton joint from time series prediction model,this study has established the minimization of human-exoskeleton interaction control strategy consist of the force feedback inner loop and the PID position tracking outer loop based on the admittance model.The simulation model was built by MATLAB,which verified that the control model can drive the exoskeleton and track the human body's movement in real time,and effectively control the mutual interference between human and machine in the motion process.Finally,the study has built up a exoskeleton robot experimental platform named as ESR-I.Based on the ESR-I platform,we carried out the experiment of the system's about its basic parameters of load carrying capacity,and the effectiveness of the perceptual control algorithm and strategy designed in this paper was demonstrated.Then with the human-exoskeleton interaction force as the index,the design experiment shows that the control following quality and compliance of the admittance model control algorithm designed by this paper is obviously improved compared with the joint torque compensation control method.Then this study has designed the experiment to measure the assisting efficiency of the exoskeleton under static standing and walking conditions based on two objective indicators of plantar pressure and shoulder strap tension.The result demonstrates that the ESR-I exoskeleton robot can effectively assist the human body to reduce the weight of the operator during the load-carrying exercise,and enhance its motion capacity. |
PDF Full Text Request