Research On The Vibration Control Of Industrial Robots With Joint Flexibility

At present,industrial robots have been widely used in automotive,military,electronics,aerospace and other industries.With the development of intelligent manufacturing technology,industrial robots are developing at high speed,high precision,heavy load and light weight.Joints of industrial robots usually compacts RV reducers,harmonic reducers,or synchronous belts as the transmission mechanism and the control system assumes that they are rigid systems.However,due to the compact structure of the robot joints,the rigidity is limited and joint flexibility is unavoidable,and vibration can be caused during high-speed motion and start-stop process,which reduces the trajectory accuracy and positioning efficiency.As a result,the vibration suppression technology of the robot's joints is of great significance to improve control performance.To this end,this thesis conducts research on the suppression of vibration caused by joint flexibility of industrial robots from three aspects: typical position closed-loop vibration suppression technique,tip-acceleration feedback vibration suppression technique,and trajectory optimization technique.The main contents are as follows:First,based on the Lagrangian energy method,a complete dynamic model and simplified model of the flexible joint robot are established.The vibration characteristics of flexible-joint robots are analyzed.After that,the conventional dynamic surface control methods are theoretically analyzed and simulated.The results show that,due to the coupling effect of the observer and the control part of dynamic surface methods,the parameter tunning process is difficult.In addition,the velocity estimation of current observers is realized by first-order filters,which leads to the weakness in anti-noise capability.Therefore,novel kind of observers are necessary.Based on the principle of error saturation technology,error saturation functions are utilized to replace the first-order filters of traditional observers to improve the anti-noise capability of the observers.Because the motor-side model of robots are easily obtained,the uncertainties of the model are concentrated on the link-side part and joint stiffness.The stability of the observer is analized by the Lure-type Lyapunov functions.After that,the observer is combined with the dynamic surface method and simulation analysis is performed on the joint unit of the robots.The results show that the method has stronger anti-noise performance and is eaiser to tune.However,dynamic paramters are difficult to be obtained for some industrial robots.In this case,based on high-order sliding-mode technique,a modified extended-state observer is proposed to improve the anti-nosie and adaptive ability.The stability of the observer is proved by using Lyapunov methods.Then,the observer is combined with dynamic surface method and simulations are performed on a robot.The results show that the method has stronger noise resistance and approximation of uncertain terms.In addition,most industrial robots can't obtain link-side vibration informations because they are only equipped with encoders on servo motors and not equipped with encoders or joint torque sensors.In the thesis,a low-cost,small-volume,easy-to-install accelerometer is used to solve the problem.A new type of end acceleration feedback method is proposed to solve the problem of uncontrollable error amplification of the existing acceleration feedback methods.Based on the modified robot model,a BP neural network is used to design a vibration suppression controller,and a flexible joint robot is used for simulation analysis.The results show that this method has good vibration suppression performance.In addition to the above two online strategies,an offline vibration suppression strategy based on trajectory planning methods is studied to further suppress the vibration of conventional industrial robots.A decomposition and reconstruction process is added on the basis of the shaping method,and the shortcomings of large trajectory distortion and poor compatibility with different trajectory optimization methods are solved.Based on the new trajectory optimization method,simulation analysis is performed.The results exhibit that the method greatly reduces the distortion of the original trajectory and indicate that this method has good compatibility.Finally,a PC-based control platform has been built using Twin CAT(The Windows Control and Automation Technology).The vibration suppression methods described above have been applied to a six-degree-of-freedom robot and a homemade two-degree-of-freedom robot to conduct experimental research and verify the proposed vibration suppression controller.The simulation results have a good agreement with simulations,and exhibit the effectiveness and practicability of the proposed methods.