Research On Motion Control Of Robotic Manipulator Based On Visual Perception

With the rapid development of science and technology and modern industry,robots play an increasingly important role in various fields.Ability to adapt and perceive the environment of the traditional robot is a little poor which can not meet the needs of mankind.In recent years,visual technology has a rapid development,machine vision prompt robot become more intelligent to adapt to environment and the visual system and technology of visual perception has become an important part of intelligent robots.The robot with visual perception has become a hot research direction for social scientific researchers in nowadays.The object fetching system of binocular vision robot has been deeply theoretical studied and experimental verification.Based on the Linux development environment,the object fetching system of robot with binocular vision perception is designed and implemented.The embedded development platform obtains the spatial coordinate information of the target object by processing the target object image acquired by the binocular camera and transmits it to the controller of manipulator.The robot control mechanism controls the joints rotate to fetch the target object according to the adjustment strategy of arm.The main contents of this thesis is divided into the following three aspects:(1)Research on the hardware which comprise of the robot and complete the design of robot structure.It mainly includes the research of the control principle of each module of the robot and the internal and external parameters and the distortion coefficient of the camera by camera calibration.(2)The camera imaging model is studied to complete the binocular depth measurement and the relationship of Eye-in-Hand calibration.Then,the three-dimensional coordinates of the target object relative to the robot gripping mechanism is obtained.Firstly,the camera is calibrated by the principle of epipolar constraint,and the SGBM algorithm is used to stereo matching the captured images.The three-dimensional coordinates of target object in space is got after the depth recovery and Eye-in-Hand calibration.(3)Research on the forward and inverse kinematics of the robotic gripping mechanism,designed the automatic adjustment strategy of manipul ator joints,and verified the structural rationality of the manipulator.