Arms' Collision-free Path Planning Of Mechanical Dual-arms Robot With Vision

The application of mechanical dual-arms robot has enhanced the robot's adaptability of complex tasks and improved the utilizable rate of robot's working space. The research of mechanical dual-arms'harmonious movement has become a chief orientation of robot in further developing.This paper has taken the AS-Mrobot and ASR robot with vision which were produced by Guang maoda partner robot limited company of Shanghai as study object, built image processing algorithms library of image acquisition card with Visual C++program software, achieved the object recognized automatically and the collision-free path planning of dual-arm robots when they are moving, and solved the problem of the mechanical dual-arms' harmonious movement, avoided collision between two mechnical arms. All of these can establish a basis for mechanical multi-arms'movement.By the comparison of some image processing methods, such as image preprocessing, image segmentation, corner detection and image matching, the better algorithms for the system have been found and applied to get the image coordinate of camera calibration model and recognize target. The experimental results show that the mechanical arm could find the object automatically when the object template having been built up and actualized non-intervention grasping movement.By studying the camera model of linear and nonlinear, a camera model with one step radial distorted has been built. Considering the fixed position between camera and the mobile robot in the experimental equipment, a new camera's calibration method which is between traditional calibration and self-calibration method are presented in the thesis to calculate the camera's intrinsic and extrinsic parameters. It is so simple that it only need to know the data of world coordinate and computer image coordinate and needn't to ascertain the exact camera's model and its interior and exterior parameters.By studying the foundation of time-varying C-space and the collision detection based on false measure distance function, time-varying C-space of slave arm has been built. Ant colony algorithm has been used to plan the collision-free path for slave mechanical arm in C-space and primary effect has been achieved. Because of the limitation of the basic ant colony algorithm for this system, and in order to improve the search speed and accuracy, we have studied on the method of compartmentalizing reseau and making the system further searching the optimal path when there are multi-feasible paths or non-feasible path. The experimental results show that the improved ant colony algorithm is better than the basic one in the global searching performance and searching speed, and can achieve the function of collision-free movement between the two mechanical arms.The methods of image processing, camera calibration and path planning that mentioned in this paper can avoid the collision in the movement of the dual-arms, and enlarge workspace. It's simple to control the dual-arms with these methods as well. With certain practical application values, these methods can be applied to the fields related to the production of mechanical arms.