Integration of vision/force robot control for transporting different shaped/colored objects from moving circular conveyor

This paper will propose algorithms for transporting different objects from moving circular conveyor with the help of vision/force robot control. Poses and speeds of the objects are calculated from sequential images captured by 3D-camera mounted on the hand of the robot. In the conveyor there are 4 slots which contain four differently formed and coloured objects. The cross sections of these objects are the square, the rectangle, the triangle and the circle. The required object will be segmented and then recognized with the help of fast image processing algorithm depending on its colour and shape. Furthermore, this paper will improve algorithm for grasping and pulling-out the required object from slot with tight clearance. For this task two different kinds of simple force control have been developed. The objects will be grasped by a 6-DoF robot with a primitive two-finger-gripper as end-effector. In the first experiment the proposed system will follow the required object until the movement of the conveyor stops. After that it will start to pull-out the object. In more complicated case, the robot will pull-out the object while the conveyor is moving. In other words, the movement of the robot will be synchronized with the movement of the object and it should grasp it at a convenient time and pose without any prior knowledge about its velocity. This paper will be supported with real experiments presenting promising results which illustrate the capability of grasping a stuck moving object with the help of vision/force robot control.