Research On Identification And Optimization Of Dynamic Parameters In The Physically Feasible Region Of UR10e Robot

Industrial robots are an important manifestation of measuring the level of industrial automation in a country.Industrial robots are usually widely used in power,logistics,chemical and other industrial fields,saving labor and production costs,but also speeding up the process of industrial automation.For some high-precision fields,people have put forward higher requirements for the motion performance of industrial robots,such as dynamic response,system robustness,positioning accuracy,etc.A hot issue in the current industrial robot field is how to achieve high-precision motion control.Generally,the composition of an industrial robot includes a mechanical body,a drive system,and a control system.These three major parts determine its highly complex electromechanical and mechanical characteristics.There are inevitably joint coupling,time-varying parameters,and load changes,which cannot be modeled.Uncertain factors and unknown external disturbances have largely affected the dynamic control effect of the system.This thesis focuses on the goal of improving the dynamic performance of the industrial robot servo system.Considering that in the actual dynamic operation of the robot servo system,the trajectory tracking accuracy based on the model motion control strategy is not high and the dynamic performance is often poor due to the inability to accurately obtain the relevant dynamic parameters.Research on the establishment of industrial robot dynamics model,parameter identification and optimization.The main contents are as follows:The research of industrial robots is inseparable from the establishment of a basic mathematical framework.This thesis first starts with the coordinate transformation of the object coordinate system in the three-dimensional space relative to the base coordinate system,and uses the description method of the DH coordinate system to analyze and solve forward and inverse kinematics,finally the space transformation matrix of a typical six-axis UR10e industrial robot is represented by symbols,and the transformation matrix is inversely calculated by analytic method to obtain the symbolic expression form of each axis angle of the joint space.In order to simplify the model establishment and verification process,the solution process uses Peter Corke’s robot toolbox to verify the results of forward and inverse kinematics.In order to achieve high-precision motion control of industrial robots,the establishment of dynamic models is an indispensable part.Common robot dynamics parameters can be simply recorded through the URDF file.Considering that the industrial robot is running in actual work,the main body linkage will have problems such as centroid offset and inertia misalignment due to the existence of other driving structures.Starting with the drive structure of the joint powertrain,a servo motor-reducer-link cascade equivalent model is constructed.The NewtonEuler algorithm is used to derive the dynamic expression form of the multi-axis industrial robot servo system,and the related dynamic parameters are corrected.Based on the ideal parameters given in the initial URDF file,design an independent servo control system for each joint of the industrial robot,and finally verify the feasibility and effectiveness of the established dynamic model of the industrial robot servo system through simulation.The control strategy based on dynamic model design has outstanding advantages over pure kinematics multi-closed-loop PID control,but the model-based control strategy relies heavily on the accuracy of internal parameters,so it is of great significance to identify and obtain the minimum set of dynamic parameters from established equivalent dynamic model.The basic idea is to linearize the established dynamic model first,remove the linearly related dynamic parameter terms,and reorganize them into the combined coefficient form of minimum set of dynamic parameters;then select the trajectory excitation with the optimal condition number after optimization.Data acquisition of torque,joint position,velocity,acceleration,and filtering process;finally,the minimum dynamic parameter set is identified through WLS in one step.In this thesis,the open source OpenSymoro software is used to linearize the industrial robot dynamic model,which greatly simplifies the parameter identification process.The system robustness of the dynamic parameters obtained by the classical identification algorithm is weak,especially under the premise that the friction force cannot be accurately modeled,the least square solution obtained through the parameter identification may not meet the requirements of physical consistency.This thesis chooses to use a improved nonlinear friction force model to compensate for joint friction,derives a unified expression form for the physical consistency of dynamic parameters,utilizes the characteristics of SDP convex optimization algorithm that the local optimum is the global optimum parameters,identifies and optimizes the minimum dynamic parameter set within the feasible region,and obtains the optimal parameter with the smallest linearization expression error from the dynamic model.By comparing the parameters obtained from the previous WLS identification,the performance superiority of the SDP optimization algorithm can be verified in the feasible region.When there are many joints,the complexity of the system increases.At this time,independent joint control often fails to achieve good dynamic results.This thesis performs dynamic feedback linearization processing on the entire UR10e robot controlled object,and uses a model-based feedforward compensation control strategy.The priori torque compensation of the controller effectively improves the dynamic tracking speed and dynamic accuracy of the system.Finally,the open source robot simulation software called CoppeliaSim is used to achieve high-precision trajectory tracking control applications.