Design Of A 6 DOF Manipulator System For Autonomous Following Human Arm Motion

The general human-robot interaction methods mainly include offline teaching and online teaching with special tools.After these methods,the manipulator can only move repeatedly according to a fixed trajectory.With the development of human-machine collaboration technology,the motion trajectory of the robot arm is often taught online according to the needs of the working conditions,so as to give play to human intelligence and the load-bearing capacity and accuracy of the robot arm.so,this paper designs a 6 DOF manipulator system that autonomously follows the motion of the human arm,and realizes two control algorithms that mimic the similar motion of the human arm and the end of the manipulator accurately follows the motion of the end of the human arm.The specific work content is as follows:(1)This paper reasonably designed the mechanical structure of the 6 DOF manipulator and the special harmonic reducer at the end of the manipulator.At the same time,the Kinect human arm bone joint data acquisition and processing software is designed on the upper computer;the software and hardware system for controlling the movement of the mechanical arm is designed on the lower computer.(2)Designed the Kinect bone data collection and smoothing preprocessing algorithm.After the bone node data is collected,the bone bitmap is depicted,and a reasonable recognition method is designed for the specific operating human body.Based on the analysis of the original bone node data information characteristics,a limited Holt dynamic secondary is designed The exponential smoothing filtering algorithm effectively eliminates the interference data of the bone nodes.(3)Established a 6 DOF mechanical arm to simulate a similar follow model of human arm motion.After transforming the skeletal coordinate information described in the Kinect coordinate system to the human two hip coordinate system,a 6 DOF joint mapping algorithm based on the joint space vector method is designed.At the same time,the joint difference between the robot arm and the human arm and the Kinect For the detected blind area of occlusion,the joint angle boundary and special position restriction algorithm are designed.(4)The trajectory model of the end trajectory of the robot arm is established to accurately track the end trajectory of the human arm.After establishing the multiple coordinate transformation model from the Kinect coordinate system to the manipulator coordinate system,the forward and inverse kinematics model of the 6 DOF manipulator with the first three joints intersecting at one point was established based on the spin theory.The angle-raising method effectively simplifies the inverse solution process.The example verification of the forward and inverse solution shows that the spinor method and its improved algorithm are correct and highly accurate.At the same time,the Jacobian matrix of the 6 DOF manipulator is solved based on the screw theory,and the singular configuration of the constitutive configuration is solved.(5)In order to eliminate the jitter of the manipulator,a disturbance-free continuous smoothing algorithm for the joint space is designed.On the basis of analyzing the characteristics of the original angle data,the third-order Bezier curve smoothing algorithm is adopted,and the smooth splicing of multi-segment Bezier curves is realized.The experimental results show that the 6 DOF manipulator can adapt to the arms of different body types,imitate the motion of the human arm with a high degree of similarity,or track the end trajectory of the human arm with high precision.