versión On-line ISSN 0718-0764
PENA-CABRERA, Mario; LOPEZ-JUAREZ, Ismael y RIOS-CABRERA, Reyes. Learning Process using Robust Algorithm to Obtain Object POSE on Real Time Assembly Lines with Robots.. Inf. tecnol. [online]. 2006, vol.17, n.2, pp.61-69. ISSN 0718-0764. http://dx.doi.org/10.4067/S0718-07642006000200009.
In this paper, the methodology and algorithm used to obtain the POSE (position and orientation) of real objects on robotic assembly lines is showed as a practical case on industrial environments. The POSE information and the features of the object are integrated in a descriptor vector [CFD&POSE] to be used with a FuzzyARTMAP artificial neural network to learn about assembly workpieces and its location. A learning process of the location of the parts for future use by the robot is obtained. The paper shows the algorithm used and the simulation done in MatLab 5.0, with concrete cases of simple parts. It also shown the experimental results obtained in a manufacturing cell with a Kuka R15 robot of six degrees of freedom.
Palabras clave : learning process; robust algorithm; assembly line; robot; real time.