Research On The Pose Estimation Of The Grasping Manipulator Based On Machine Vision

In the context of increasingly fierce industrial competition,manufacturing in2025 is becoming more and more important.Industry 4.0 has also become China's development goal in recent years.As a representative field of industrial development,smart manufacturing is increasingly valued by the state.The degree of automation in the manufacturing industry largely reflects the development level of a country.Therefore,more flexible manipulators are used in industrial production.More and more complex tasks require robots to complete independently,and the robot arms can independently complete the tasks assigned to them.The fixed tasks,ranging from the processing of aircraft and ship parts to the tightening of the screws of mobile phones and computers,are inseparable from these robotic arms.Therefore,more and more automatic robotic arms are being put into use to make them replace manual,but in When the external environment changes,they cannot make corresponding countermeasures,so this type of robot can no longer meet some specific work scenarios.In order to cope with these problems,the visual robot was designed to solve most of the problems of the traditional robotic arm.The core problem of the machine vision servo system is pose estimation,to solve the problem of pose estimation,to improve the acquisition of the robotic arm The ability of external information and the ability to respond are of great significance.In this paper,the hybrid vision manipulator on the production line is the main research object,the problem of pose estimation is studied,the kinematics model of the manipulator is established,and a new EKF/UKF-SVSF algorithm is proposed to make the manipulator more competitive.High pose calculation accuracy and robustness;for information collection,on the basis of monocular vision,a multi-camera configuration system is proposed,using a distributed data fusion algorithm based on OWA operator,which greatly increases the information collection Accuracy,the specific work is as follows:(1)The kinematics analysis of a six-degree-of-freedom robotic arm was conducted.The geometric relationship between the link and the joint of the robotic arm was studied,and the forward and inverse kinematics of the robot were solved.The process is actually a valid demonstration of the optimal solution for the motion path of the robot arm.Then,the article uses the Robotics Toolbox in matlab to simulate and verify the algorithm used in the above solution.(2)The imaging principle of the camera is briefly described.A camera model is created by using the rcvtools tool in matlab.The camera model and the robot armmodel are combined to perform the overall modeling and simulation of the robot arm visual servo system.(3)For the pose estimation of the robotic arm,the traditional Kalman-based filtering algorithm and the smoothing variable structure(SVSF)-based filtering algorithm are introduced.Combining these two algorithms,a Kalman smoothing method combining both is proposed.Variable structure algorithm(EK / UK-SVSF)estimates pose.(4)Compared with the traditional monocular vision robot,this paper uses the method of multi-camera fusion to achieve accurate estimation of poses.A distributed data fusion algorithm based on ordered weighted average operator is proposed.The comparison of simulation results proves The feasibility and advantages of the algorithm.