Research On The Design And Control Strategy Of Hip Joint Power-assisted Exoskeleton

Exoskeleton robots are widely applied in medical rehabilitation,exercise enhancement,exercise assistance and other fields,which also contain numerous research hotspots.This article aims to accurately apply the desired assist force and reduce the metabolism of healthy people,who is the target user group.Through analysis,the flexible exoskeleton of the hip joint is selected as the research object and the corresponding exoskeleton system is designed.The main research content includes hardware,software,simulation platform design and also the upper and lower controller design.First of all,analyze the current status of exoskeleton research,and combine physiology,human kinematics and other disciplines to study the human body’s movement mechanism and gait law.It is believed that the design of the hip joint flexible exoskeleton robot can achieve better assistive effects.Comprehensive analysis of the limiting factors and research purposes of the use scene,a set of hip exoskeleton prototype system is developed,which mainly includes a hardware platform,host computer interactive software,simulation platform and corresponding control system.Secondly,study how to obtain the posture information of the target joints repeatedly and accurately through IMU.The posture information is an important sensor signal of the upper and lower controllers.For this reason,methods to eliminate IMU binding error and IMU attitude calculation are discussed.After giving the relevant posture expression method and describing the definition of the coordinate system,the improved joint correction algorithm effectively reduces the IMU binding error.On this basis,a method for obtaining the a priori optimal solution parameters is proposed.The parameters are used in subsequent pose calculations.By comparing the accuracy of various calculation algorithms and the ability to resist magnetic field interference,the two-stage Kalman filter algorithm is finally selected to fuse the corrected IMU measurement data,so as to accurately obtain the hip joint posture information.Furthermore,the design of the upper-level controller is discussed,the purpose of which is to generate the corresponding ideal auxiliary force at the specified gait position,and the force can be designed for different motion modes.To this end,the adaptive Hopf oscillator based on Dynamic Hebbian Learning is first used to learn the dominant frequency of limb movement to achieve accurate calculation of real-time gait cycle information,and then the two motion modes of walking and running are realized through the LDA pattern recognition algorithm under a specific combination.With the above-mentioned real-time gait cycle and motion mode calculation results as input,the ideal auxiliary force output is realized through the polynomial interpolation method,and the final pattern of force can be adjusted through 6 related parameters.Finally,the design of the lower-level controller is explored,aiming to accurately realize the ideal auxiliary force generated by the upper-level controller.A variety of control methods are designed for the two application conditions,and the effectiveness of each method is horizontally compared through simulation experiments and actual experiments.On the basis that the lower controller can better track the ideal auxiliary force,and the upper controller adjusts the control point of the ideal auxiliary force,the research goal of comfortably assisting human body movement and reducing sports metabolic energy is achieved.