Research On Integrated Joint Torque Sensing And Compliant Control Of Collaborative Robots

With the continuous development of the manufacturing industry,there are more and more task scenarios that require human-robot collaboration,which puts higher requirements on the intelligence of robots.Collaborative robots are borned.Collaborative robots have two major advantages over traditional industrial robots.They can realize external force sensing and active soft control at the joint level.In this thesis,we focus on the external force sensing and compliant control of the integrated joints of collaborative robots.The main research content are as follows:With the premise of cost saving,a joint moment estimation algorithm is proposed based on the harmonic transmission principle of the integrated joint of the collaborative robot without using the joint embedded moment sensor.Firstly,the harmonic transmission mathematical model of the integrated joint of the collaborative robot is established based on the structure of the harmonic reducer.Then the adaptive Kalman filter mathematical model is designed based on the Kalman filter principle combined with robust filtering.Finally the simulation model is built in Matlab/Simulink to verify the correctness of the designed algorithm.Although the observation of joint moments can be achieved without the application of a moment sensor,its accuracy is inferior to that of a joint moment sensor.Therefore,this thesis proposes the design of a new type of torque sensor.Firstly,after determining the design objectives based on the requirements,the design scheme of the new torque sensor is proposed.The 3D model of the joint torque sensor is designed in Solid Works and the finite element analysis and topology optimization are performed by using ANSYS.Then the deformation measurement system of the sensor is designed,and finally the whole sensor is fabricated and installed.The experimental tests are conducted on the calibration test bench.Finally,the linearity is 3.26% and the sensitivity is 1.02V/N·m.Commonly used human-machine interaction robots usually can only achieve the interaction function between the end of the robot and human.This thesis investigates the compliant control algorithm at the robot joint level.In this thesis,a zero-force control algorithm based on moment estimation for dragging demonstration scenarios,a position-based impedance control algorithm for scenarios with high force control requirements and an adaptive force tracking control algorithm improved on the basis of impedance control are proposed.Model simulations were performed in Matlab/Simulink to initially verify the feasibility of the algorithms.Finally,the experimental platform was built with the integrated joint designed in our lab.The test of the zero-force control algorithm based on moment estimation and the adaptive force tracking control algorithm based on impedance control designed in this thesis were completed on the experimental platform.The experimental results show that the designed flexible control algorithm can basically achieve the expected functions.The tracking delay is less than 0.2s and the tracking error is less than 0.5 N·m.