| Περίληψη: | Robotic applications in industry today are supported by a multitude of sensors. Sensors are used for many applications such as autonomous navigation, to counter shopfloor uncertainty, for object detection, human-robot-collaboration, or to measure the amount of grip and pressure required to hold an object. Prevalent among the various types of sensors are the visual sensors, cameras. When mobile robots are used in industry, in order to move from workstation to workstation with accuracy, usually physical docking methods are used, which add physical constraints to the system. While image processing is capable of supporting its own applications, it can also be used to improve upon the uncertainty of applications caused by other sensors, thus making physical docking obsolete. In this way it can be a key component of reconfigurable systems. Current research considers navigation and localization mostly as the same module, while object detection is researched completely separately. In this thesis, methods for autonomous localization and object detection for compensation of the error accumulated from the autonomous navigation of a mobile robot through the use of image processing are presented, as well as how they are integrated on the overall robotic system. These methods are considered parts of a corrective framework which provides the necessary accuracy to handle objects without the need of rigid restraints.
The validation of the thesis was performed using a Mobile Robot Worker (MRP), with on-board vision sensors, in an automotive use case that focuses on reconfigurability. The automotive use case focuses on an assembly line of an automotive company producing the front suspension of a passenger vehicle. In this study, the Mobile Robot Worker is used to pick and place objects in three workstations, as well as perform screwing operations while the workstation and the robot are moving in parallel. The framework introduced in this thesis was validated on the localization of the MRP in each station and the handling of objects in those stations.
|
|---|
