Intelligent Control Of Robot Arm Based On Three Viewpoint Calibration

In our society, new technology has developed rapidly. Big data, artificial intelligence, robot learning and in-depth study have become a hot research topic. At the same time, it also promotes the development of the robot toward the direction of intelligence. Also, to some extent, the level of machine intelligence reflects the level of a country's industrial modernization. Machine intelligence is also an interdiscipline, which combines artificial intelligence, in-depth study, sensor technology and motion control theory. Manipulator arm as a main research direction of machine intelligence has a wild range of applications. Currently, programming sets most manipulator arms, which mean that these manipulator arms repeat the same thing under a certain condition. Most manipulator arms do not have the perception of external environmental information and cannot make the related determines to do the related homework. With the rapid development of industrialization, there are more demands of intelligent manipulator arms, such as handling manipulator arms, welding robots and the use in the space, which all need manipulator arms to sense the surrounding environment information, to process related motions by using captured information. The development of intelligent manipulator arms is based on computer vision technology, so when computer vision technology has a breakthrough, it will promote the development of intelligent manipulator arms.The servo control system of the robot arm based on vision can be divided into different systems according to the different number of cameras. The manipulator servo control system based on vision camera according to the number of different points for different systems can be divided into monocular vision servo control system of the manipulator, binocular vision robot arm control servo system and multicast visual manipulator servo control system. With the increasing number of cameras, the controlling process of manipulator arms will be more difficult and the time complexity of processing will be higher. Currently, most research is based on intelligent manipulator control system with two viewpoints.In this paper, a servo control system with three eyes and a fixed eye is used. Manipulator arms'operation to objects is finished by using cameras to obtain 2-D picture information in the space, then using 2-D picture information to obtain the real positions of objects in the space and the position of manipulator end control unit, last using manipulator end control unit to operate planning trajectory.Through the control algorithm, complete the mechanical arm of the object related operations.The kinematics model D-H model of the manipulator is deduced and solved in detail. In the trinocular vision, via trifocal tensors to complete the interior parameters of the camera calibration, and according to the computer vision technology,the completion of the image feature extraction and preprocessing.We use MATLAB to simulate. The accuracy of the model is verified by the simulation of the six degrees of the freedom of manipulator's kinematics algorithm. In the three interpolations and five interpolation simulations, it is proved that the five interpolations are more suitable for the kinematic control of the manipulator.