Usually, industrial robots are dedicated to one task during their lifecycle, where manufacturer provides certain sensorics and software environment, where a robot can be programed. The problem arises when we want to use a robot for advanced aplication where additional sensorics is required. ROS platform enables us to plug sensors to personal computer and to program the robot in programming languages C++ or Python. With the help of ROS personal computer advanced applications can be developed. These applications include real-time adjusting of the robot to the environment with the help of sensors.
In this thesis, model of ROS (Robot operating system) compatible robot was created. The process of connecting and calibration of additional camera is presented. Computer vision algorithm for hole and object detection was developed. Different planners available on ROS platform are evaluated. ROS service to control digital outputs/inputs, which enables us to control the gripper was developed. Finally, pick and place application was developed. Pick and place application uses successfully detected objects and holes from the camera for picking the objects and placing them in the free holes.
|
