Research On Direct Teaching Of Robot Based On Dynamic Model

With the development of “Made in China 2025”plan,the demand for intelligentization in the industrial manufacturing field is growing and collaborative robots have become a new research hotspot.When a traditional industrial robot teaches and playback in the production environment,it needs to perform programming operations on the teaching box for each action and perform programming training for operators.Therefore,direct teaching has become a new research hotspot,which has the advantages of convenience and simplification.Under this background,Some research on the direct teaching of industrial robot based on dynamic model is made.There are many ways to realize direct teach.Dynamic model includes the gravity,friction inertia force is established and then external force is calculated based on dynamic model to realize direct teaching.In order to realize the research of direct teaching of collaborative robot based on dynamics model,firstly,an experimental platform is built based on the existing resources in the laboratory.Effort's ER3A-C60 robot is tabken as the research object.The controller adopts VxWorks operating system.The EtherCAT communication protocol is adopted between the controller and the servo driver.Based on the multi-task mechanism of VxWorks,real-time sampling of the robot's running state and feedback control in position mode are realized at the same time.According to the positive relationship between the output torque of the motor and the current,the moment of the joint during the operation of the robot can be sampled.The software framework design and implementation of the collaborative robot teaching process and the various software modules in the process of reproduction.Then,the dynamic model of the robot is analyzed.The dynamics models such as gravity,friction and inertia of serial robots are modeled according to Newton's Euler method.There are 12 parameters to be identified of each axis.The relationship between the parameters to be identified and the output torque of the joint space of the robot is analyzed to be linear.Based on the condition number,the error sensitivity of the coefficient matrix to the sampled data is analyzed to generate a Fourier-type excitation trajectory.There are many parameters to be identified in the dynamic model of robot based on Newton-Euler method.The deviation of each parameter may lead to the inaccuracy of the model.The neural network is applied in the dynamic model.The neural network can predict the dynamic model due to the universal approximation theorem.The open source database Tensorflow is used to establish the neural network.And then the paper compares neural network and analyzes the dynamic model established by Newton Euler's method.Finally,the established dynamic model of neural network is used to analyze the direct teaching of the cooperative robot.In the study of drag and teach,the follow-up motion of the co-robot to the external force is realized by.Physical verification is performed in single axis and then performed in the unity of the first three axises.Through experimental verification,the use of neural network dynamic model can be a good realization of the robot's direct teaching.