Research And Implementation Of Binocular Mobile Robot And Autonomous Grasp Of Robot Arm

With the development of robot related technology,its intelligence has been increasingly concerned.However,current mainstream robots still repeat the specified actions according to the program,and do not have the ability to obtain outside information.Once the position is changed,the robot cannot accomplish the task,its intelligence needs to be improved.Thus,people shift their eyes to the field of machine vision.The binocular vision technology can simulate human eyes to obtain rich information from the outside world,and use triangulation principle to obtain depth information.So the introduction of vision in robots has become a research hotspot for improving the intelligence of robots,and it also provides a basis for solving robot autonomous control problems.The research work of this paper is mainly divided into two parts.One is the robot motion system,the other is the robot vision system.For the motion system,the paper designs and constructs an omni-directional mobile robot.The design work is performed from two aspects of mechanical structure and control system.The mechanical structure mainly includes the omni-directional mobile platform and the six degree of freedom manipulator.The control system includes hardware circuit design and programming.Then the paper analyzes the kinematics of platform and manipulator,and the rationality of kinematics mathematical model is verified by simulation experiments.For the vision system,the paper designs and constructs the binocular vision platform and gives its mathematical model.Then it studies the problem of binocular location.Firstly,the camera imaging model is introduced,and the internal and external parameters required for camera calibration are derived.Through the analysis of the traditional calibration and self-calibration method,the paper chooses the Zhang Zhengyou calibration method between the two,and uses the stereo correction to verify the validity of camera calibration results.Secondly,the image preprocessing program is determined according to the need.It mainly includes image graying and image filtering.In addition,around the stereo matching problem,the paper chooses the SAD matching method to get disparity image.Finally,for the target location problem,a localization method based on disparity geometric model is study,and uses this method to locate the target quickly and effectively in the laboratory environment.Based on the research content of above two systems,Firstly,the typical motion of the motion system is tested to obtain the motion speed;secondly,the robot autonomous grasping vision system is designed with OpenCV in Visual Studio 2010 IDE.And the experiment of ranging and target grabbing is completed by the vision system;in the end,the analysis of the above experimental results shows that the vision robot system has good accuracy and reliability.