The present work shows recent results in the robotic manipulation context, after suitable integration between computer vision and radio frequency identification (RFID). The experimental validation has been carried out in a household environment, by using current manufactured objects, labeled previously, and processed through a RFID reader situated in the robot hand. A camera-in-hand also was used for segmentation and localization of the corresponding objects. The main problems solved in this work are those related with hardware integration between the robot and the RFID system, including the RFID reader, antenna and necessary interfaces. Furthermore, a graphical user interface was developed for displaying the results from visual processing, including both, segmentation and localization of the objects in the scene, visualizing at the same time the corresponding RFID labels associated which each previously identified object. These preliminary results have demonstrated the feasibility and reliability of this system to succeed with localization, identification and handling of daily objects, without predefined models at all, in household environments.

Read More: http://www.worldscientific.com/doi/abs/10.1142/9789812835772_0101