Research On Application Of Control Technology Of Industrial Robot Arm Fusing Vision And Force Sense

Due to its high efficiency,good stability and strong adaptability to harsh environments,industrial robot arm has been transformed into productivity for social development in many fields.However,the traditional industrial robot arm also has shortcomings such as poor autonomy and low compatibility.They work in fixed parameters,processes and even fixed motion trajectories.Even if small changes occur in the process or work object,the operating system and manipulator motion parameters need to be reseted.Therefore,the intelligent research of the robot arm is a hot issue in recent years.The paper mainly studies the control technology of industrial robotic arm that combines visual guidance and force feedback.It uses visual algorithm to detect and locate the target,and uses force feedback to control the contact between the industrial robot arm and the target.The main work contents are as follows:The eye-to-hand relationship matrix of the industrial camera fixed in the working environment and the end of the robot arm is calibrated,and the object image collected by the camera is image-processed to obtain the target position.Perform preprocessing such as graying,filtering and denoising on the acquired image,using background subtraction,contour extraction and obtaining the minimum circumscribed rectangle to obtain the coordinates of the center point of the target in the image coordinate system,then transform the coordinates of the center point through the calibration matrix to the robot arm base coordinate system to achieve the target position.On the basis of visual positioning,the end of the robot arm is controlled to move to the top of the target and then slowly approaches the target.Based on the machine learning method,the force information in the process of the end of the arm contacting with different characteristic objects is analyzed.The force information sample were generated by the end of the robot arm picking up 10 types of 28 different bags of small foods,and the force information sample are used to train the LSTM network to derive the force feature detection model.The model is used to judge the characteristics of the force generated in real time during the contact process between the robot arm and the object,thereby controlling the action of the robot arm when it contacts the target.Design a robotic arm application system that combines visual guidance and force feedback.,the software system is used to control the robot arm to perform practical experiments on the picking and packing operations of specific static objects and moving objects on the conveyor belt by constructing an application experiment platform.Finally,the application system achieves that using the software system to control the robot arm to complete the picking and packing of objects placed randomly which is static or on the conveyor belt,which verifies the effectiveness of the system,and shows that the industrial robot with integrated visual guidance and force feedback is more flexible and greater intelligence than the traditional industrial robot arm.