Research On The Terrain Parameters Fusion Based Gait Autonomous Decision-Making Method For Paraplegic Walking Assistance Exoskeleton And Its System Implementation

Currently,exoskeleton robots for paraplegic walking assistance are primarily studied for gait planning in single or simple scenarios.However,there are significant challenges in terrain perception and adaptive transitional gait generation for exoskeleton robots in real-life scenarios that involve transitional terrains.This thesis proposes gait autonomous decision-making method that incorporates terrain parameters to enable autonomous walking of exoskeleton robots in the flat-staircase environment.The main research contributions are as follows:(1)To address the requirements of feasible space perception and staircase parameter detection in the flat-staircase scenario,we propose a terrain parameterization method based on point cloud data.This method utilizes depth cameras to capture color and depth images of the scene,and employs RGB-based semantic segmentation to extract feasible regions for providing gait transition signals to the exoskeleton robot.Then,by performing point cloud back-projection using the depth image of the feasible region,we reconstruct the spatial 3D point cloud of the stairs and calculate accurate stair parameters.Experimental results demonstrate the accuracy and stability of the forward distance parameter and achieve a stair parameter accuracy of over 94%.(2)For the transitional gait issue in the flat-staircase scenario,this thesis proposes an autonomous gait parameter decision-making method based on terrain environment.By capturing the motion state of the human-robot system and integrating perception information,Stage parameterized representation of walking gait.By retaining multiple demonstrated trajectory characteristics,this method plans personalized gait trajectories suitable for both human and environmental parameters through parameterized dynamic motion primitives,thus achieving autonomous gait decision-making.Analysis of walking with variable step lengths near the stairs and crossing stairs illustrates the completion of trajectory planning in the transition stage,transitional stage,and stair climbing stage.(3)Exoskeleton robot system setup and method validation.Based on the platform of paraplegic walking assistance exoskeleton robot,this thesis constructs a human-robot gait autonomous decision-making system that incorporates environment perception and conduct feasibility and effectiveness experiments to verify the proposed gait decisionmaking method.By comparing the natural gait trajectories of healthy subjects in the flatstaircase scenario captured by an optical motion capture system with the parameterized gait trajectories generated by the exoskeleton human-robot system,this thesis validates the effectiveness of the proposed gait autonomous decision-making method.This thesis proposed gait autonomous decision-making method can effectively enhance the intelligence and human-robot interaction experience of the paraplegic walking assistance exoskeleton system,laying a technological foundation for the practical application of walking assistive robots in the daily lives of paraplegic patients.