| Press manufacturing is an important metal forming technique in modern machining and fabrication process. Being of the advantages of rapid, efficient, material and energy economy and less pollution, press manufacturing technique is widely used in many industry fields and is developing towards precision, high-speed, automation and flexibility with the progress of Advanced Manufacturing Technology. It has been a significant trend for modern press manufacturing pro-cess to be automatic and flexible in forms of automatic press manufacturing unit or automatic production line employing automation, robot and servo technology.It is an important developing direction that the robot is accompanied by vi-sion so that it can interact intelligently with its working environment. Generally, the system of visual servoing can be divided into following subsystems:camera, image preprocessing, image understanding, visual servoing controller of robot. To improve the accuracy and range of application, camera calibration, feature detection and designing of visual servoing controller are deeply researched.Based on a visual servoing of stamping material feeding robot SR09, this dissertation researched a series of issues of robot including conceptual design of robot, kinematics, dynamics, visual servoing system as well as design of controller. The main points and fruits are following:The first chapter mainly dealt with the current trend in stamping robot and the technique of visual servoing. This chapter began with the trend in press process, then discuss of the impact of robot on press process. Also the category and configuration, coordination with press, and its synchronization and coordi-nation with whole automatic press production line are explored, analyzed and summarized. In addition, a survey of visual servoing and its related technologies are presented. It is concluded that it is highly significant to research and de-velop new generation of intelligently autonomous robot for automation of press process, enhancement of capability and level of manufacturing, enhancement of rapid response to market.In chapter two, after the components and advantages of the uniform mod-eling language (UML) being researched and analyzed, a model of modern design process based on UML (uniform modeling language) is proposed, subsequently the procedure and object of conceptual design of mechatronic system were ex-pounded. In the ending of chapter two, the conceptual design of SR09 was carried based on uniform modeling language (UML) by means of case driven methodol-ogy, and the results of simulation of kinematics and dynamics of robot SR09 was given.Chapter three based on the articulate robot (manipulator for short), view-ing the links which composes the manipulator as rigid body, applying principle of rigid body mechanics, expatiates the basic motion description and computa-tion of kinematics and dynamics of robot SR09. In this chapter, motion frame of links and Denavit-Hartenbcrg notation as well as motion transform matrix of SR09 are established; the transformation among vision frame, base frame of robot SR09 and workpiece frame also discussed. Moreover, equation of kinematics and dynamics of manipulator of SR09 are deduced and closed form solution to these equations is given; a simple and specific dynamical algorithm based on Newton Euler equation using MapleTM is proposed too.In chapter four, article concerns with methods of computing a trajectory in multidimensional space which describes the desired motion of a manipulator. Here, trajectory refers to a time history of position, velocity, and acceleration in joint space or Cartesian space. The contents includes several typical method of trajectory generation in joint angles such as Cubic polynomials, Cubic polyno-mials for a path with via points, higher order polynomials, linear function with parabolic blends, B-spline. In Cartesian space scheme. Chapter four also relates the mostly used trajectory in stamping robot, and a optimal algorithm comput-ing the work position of robot with minimum joint angles in non-singularity work space. A comparative Simulation is presented to verify a integrated trajectory that is composed several straight line and a circle in ADAMSTM at the end of chapter, and the simulation results shows that robot SR09 works well with the trajectory.Chapter five discusses the structure of vison system, pin-hole camera prin-ciple, camera calibration. Machine vision system mainly consists of acquisition of image, and process of image. Image process aim at canceling noise by digital filtering, converting color or grey image into binary image, getting the contours of workpieces through edge detecting, computing the pixel coordinates by seg-menting image. The camera calibration of SR09 is given and an algorithm of object workpiece recognizing based on BP neural network is advanced, and the algorithm verified by test.Chapter six discusses the monocular and binocular visual servoing systems and position based fuzzy PID controller. In this chapter, the principle, configura-tion of monocular and binocular with eye-in-hand are researched; including image Jacobian matrix, combination of multi-feature points Jacobian matrix, control al-gorithm of manipulator based on PD (Proportional derivative) control, binocular stereo visual servoing control system and related simulations. Because the visual servoing robot takes the vision system as the motion sens-ing unit, undergone all kinds of disturbance, characterized with hyper non-linear and time varying, the precision of robot controller is deadly affected by the en-vironment. In order to improve the autonomy and adaptability of robot and to enhance its performance, in this chapter, a fuzzy PID controller is designed for SR09. making the control parameters be modified with error and derivative of the error and the control performance be improved continuously. The tracking simulation of a circle on a slant plane verified the controller working well.Chapter seven is related to information fusion of robot. Information fusion is a process to use a numbers of synergistic sensors and combine these sensors effectively to form sensing system of high performance to acquire a consistent description of the sensed system. For the time being, the main methods to fu-sion information are weighted mean, Shafer-Dempster evidence inference, product ruler, fuzzy theory and neural network.Stamping robot contacts the environment frequently when they works in a loop of grabbing objet, transporting the object and laying down the object. It is imperative that the normal force between the end-effector and environment be maintained in a certain controlled range. When the force sensor is absence, it is possible to know the force exerting on environment by measuring the current that flow through coil of the joint motor. Therefore, a sense of access is obtained. This is a good method to control the robot to position precisely with low cost. In this chapter the principle of hybird control of force and position was analyzed with no force sensor just by measuring the current of joint motor. By means of Robotics TOOLBOX for MATLAB developed by Doctor P. I. coke,a simulation system is build up and the simulation carried out.Chapter eight is the conclusion. In this chapter some conclusions are made and a plan is advanced for future works. |
PDF Full Text Request