Resumen de Control of a Robotic Arm for Transporting Objects Based on Neuro-Fuzzy Learning Visual Information
Juan Hernández Vicén, Santiago Martínez de la Casa, Carlos Balaguer Bernaldo de Quirós, Juan Miguel García-Haro
New applications related to robotic manipulation or transportation tasks, with or without physical grasping are being developed. To perform these activities different kind of perceptions are need. One of the key perceptions in robotics is vision. However, camera-based systems have inherent errors which affect the quality of the information obtained. Image distortion slows down information processing and defers data availability to last processing stages, decreasing performance. In this paper, a new approach to correct diverse sources of visual distortions on images in early stages of the data processing is proposed.
The goal of the proposed system/algorithm is the computation of the tilt angle of an object transported by a robot. After capturing the image, the computing system extracts the angle using a fuzzy filter that corrects all distortions at only one processing step. This filter has been developed by means of neuro-fuzzy learning techniques, using data obtained from real experiments. In this way, computing time can be decreased and the performance of the robotic application can be increased. The resulting algorithm has been tried out experimentally in robot transportation tasks in the humanoid robot TEO (Task Environment Operator).
