Research On Collision Detection And Impedance Control Of Industrial Robots

With the increasing use of robots in various fields,especially for contact operations such as assembly and grinding,it has become essential to not only plan and control the robot’s trajectory,but also to control the output force of each joint or end of the robot.Accurate force detection and control can not only be used to determine whether the robot has collided with the external environment,but also make the robot show a certain degree of compliance when contacting the operation.This issue is a pressing matter in the realm of robot control,and it is a highly sought-after area of study in current research.This thesis focuses on the study of robot dynamics,specifically on a tandem industrial robot with six degrees of freedom as the research subject.The research objectives are as follows:(1)Kinematics and dynamics modeling analysis of PUMA560 robot.The kinematics model of a robot was established using the Modified Denavit-Hartenberg(MDH)method.The working space of the robot was calculated using the Monte Carlo method,and the inverse solution of the robot was determined based on the shortest stroke method.We utilized the Newton-Euler method to construct the dynamic model of the robot,and the dynamic model was linearized,and the observation matrix was extracted.Then,A linear relationship between the minimal set of inertial parameters and the actual torques was established,which establishes a basis for the later identification of dynamic parameters.(2)Robot dynamic parameter identification based on nonlinear weighted particle swarm algorithm.To achieve precise identification of all parameters,we utilized a fifth-order Fourier series as the excitation trajectory,the coefficient matrix condition is the smallest as the objective function,the trajectory coefficient is optimized by nonlinear weighted particle swarm optimization(NWPSO),and a penalty function is introduced to eliminate unqualified particles such as exceeding the joint limit.The least squares method is used to identify the minimum inertial parameter set,and the test trajectory is set in MATLAB to verify the torque.(3)Robot collision detection based on nonlinear expansion momentum observer.Based on the acquisition of the robot dynamics model,various algorithms for detecting collisions are analyzed.To address the drawbacks of complex collision detection algorithms and long detection times,a nonlinear expansion momentum observer is proposed,which considers the torque of collision as a state variable of the system,and reduces the order of the observer according to the concept of momentum,reducing the complexity of the observer to shorten the algorithm execution time.Aiming at the problem that the fal function has many parameters and the difficulty of tuning,the fal function is changed to an inverse hyperbolic sine function,which reduces the number of tuning parameters and has clear meaning.In order to improve the accuracy of collision detection of a robot,a torque detection method is introduced which considers a time-varying threshold.The method calculates the theoretical torque value of the robot based on identified parameters and uses it as a reference for detection.When the actual torque of the robot exceeds this reference value,the method determines that a collision has occurred;A collision observer model was built in MATLAB/Simulink,and simulation was carried out on a robot with two degrees of freedom and six degrees of freedom,respectively.The experimental results show that the nonlinear expansion momentum observer can quickly detect the occurrence of collision,which improves the sensitivity and accuracy of collision detection.(4)Robot compliance contact control based on adaptive variable impedance control.The concept of impedance control and how impedance parameters affect the control system were studied and analyzed.To ensure that the robot demonstrates a certain level of compliance while performing contact operations,impedance control is utilized to achieve the desired force tracking control,thereby enabling the robot to exhibit a degree of flexibility.An adaptive impedance control model is proposed,so that the damping parameters can be adjusted in real time with the change of uncertain contact force.Since the effect of impedance control requires multiple parameters to be adjusted together,this thesis introduces the variable inertia coefficient in the adaptive impedance model,and applies the impedance model to the plane,slope and surface for simulation analysis.The simulation results show that the adaptive variable impedance model can improve the compliance of the robot during contact and quickly track the expected force.(5)Co-simulation verification of MATLAB and Coppelia Sim.Using Coppelia Sim’s dynamic engine,the environment in which the robot is working was simulated,and the research content of the previous chapter was verified.