Research On The Controlling Technology Of Delta Parallel Manipulator Based On Machine Vision

Machine vision provides a means for the robots to perceive the external environment, so that the robot has the ability to locate the external things by itself. In modern industrial production filed, automated and flexible production also require the industrial robots to have the ability to sense the external environment. Delta parallel manipulator is a kind of industrial robot, which has the advantages of simple structure, flexible movement, high precision and strong bearing capacity, and has been widely used in industrial production filed, such as electronic assembly, food processing and pharmaceutical production and so on. The high speed grabbing and sorting operations for production target are required in industrial production, These operations are not easy to the human beings. In some environment, the target objects are scattered, and must be positioned them first. Considering the requirements above, a set of methods of controlling the Delta parallel manipulator based on machine vision has been studied in this dissertation. The main contents of this dissertation include the following aspectsAn optimized inverse kinematics algorithm of Delta parallel manipulator is proposed, which greatly speeds up the operation speed of the control unit. Kinematics of Delta parallel manipulator is the foundation of the motion controlling, in order to get the motion controlling algorithm of delta parallel manipulator. The kinematics analysis of Delta parallel manipulator is at the beginning part of this dissertation. The structure of the Delta parallel manipulator has been simplified in a reasonable way and the mathematic model for the movement of Delta parallel manipulator has been established in this part. On the basis of those established mathematical model, the forward kinematics and inverse kinematics of the robot arm is obtained by using the method of space analytic geometry and vector operation in this dissertation. The range of motion of Delta parallel manipulator is obtained in this dissertation by using the mathematical analysis method and MATLAB mathematical calculation software, which provides a basis for the determination of the visual range of the visual control system. Based on the existing mature inverse algorithm, a optimization of inverse kinematics solution is proposed in this dissertation by deeply analyzing the structure of Delta Parallel Manipulator, which can improve computing speed of around 10 times compared with the original solution, and can greatly reduce the computation burden of the control system and greatly accelerate the reaction speed of Delta parallel manipulator.Secondly, according to the actual production environment and the requirements of this project, a monocular and two-dimensional vision calibration method is used in this dissertation. The distortion parameters of the lens and the internal parameters of the camera are integrated into the calculation parameters, and this method is easy to operate. This calibration method is simple and rapid, and it is suitable for industrial production.Thirdly, a method of getting the center coordinates and shape characteristics of the objects by its from their images has been studied in this dissertation, According to the experimental conditions, the method of real-time tracking and dynamic grasping of objects in the moving belt has also been studied.Finally, in order to grasp the target object dynamically by the Delta parallel manipulator, a method that using software timer and speed of conveyor belt is used in this dissertation. Compared with the vision way of tracking objects, the methods mentioned above is simpler and easier to implement, In addition, it require less resource.