Research On Vision-based Special Manipulator Control

In the modern battlefield,special robots came into being in order to reduce casualties.Special robots can perform special tasks such as reconnaissance,explosive ordnance disposal(EOD),combat and more.For the task of EOD,the manipulator of special robots can achieve semi-automatic grasping and have the ability of accurate grasping.The special manipulator control system is designed with the visual servo technology,and the research is verified through experiments.(1)Firstly,the research status and development trend of visual servo manipulator are expounded.The structural design and control system are carried out by general plan.(2)According to the requirement of target location,a visual servo model of manipulator is established,and a recognition and positioning method based on monocular binocular combination is presented.When the real-time performance is satisfied,the 3D position coordinates of the target are accurately acquired.Based on monocular vision algorithm and binocular vision algorithm,compare the pros and cons of both and improve target positioning strategy by complement of monocular binocular advantage,realize the experiment of this method,verify the validity of the proposed method using the pose measurement experiment.(3)The manipulator kinematics model by D-H parameter method is established and the forward kinematics is derived and solved based on this model.The inverse kinematics solution method of the manipulator is studied,the only inverse solution is obtained through the range of actual motion threshold of the joint.For a single object and multi object grasping and placing continuous capture demand,using trajectory planning method for planning the end effector trajectory,on this basis,this paper proposes an improved algorithm of trajectory planning algorithm based on spline interpolation.Finally,use MATLAB to obtain the simulation analysis of the manipulator operation range and end effector trajectory planning.(4)The control system of the manipulator is designed.The hardware platform and interface are designed through the control process.For manipulator trajectory smoothing and changeable load demand,the joint motor fuzzy PID controller is designed.The movement status of the manipulator become stable and reliable by debugging.The grasping experiment is done and the experimental error comparison is made through the measurement results of the sensor module,which verifies the feasibility of the robot arm grasping experimental method.