Fast Object Recognition for Grasping Tasks using Industrial Robots

Reconocimiento rápido de objetos para tareas de agarre usando robots industriales

Ismael López-Juárez¹, Reyes Rios-Cabrera¹, Mario Peña-Cabrera², Gerardo Maximiliano Méndez³, and Román Osorio²

¹ Centro de Investigación y de Estudios Avanzados del IPN (CINVESTAV), México. Correo: ismael.lopez@cinvestav.edu.mx, reyes_rios@hotmail.com

³ Instituto Tecnológico de Nuevo León (ITNL), México. Correo: gmm_paper@yahoo.com.mx

Article received on 11/16/2010.
Accepted on 24/09/2012.

Abstract

Working in unstructured assembly robotic environments, i.e. with unknown part location; the robot has to accurately not only to locate the part, but also to recognize it in readiness for grasping. The aim of this research is to develop a fast and robust approach to accomplish this task. We propose an approach to aid the learning of assembly parts on-line. The approach which is based on ANN and a reduced set of recurrent training patterns which speed up the recognition task compared with our previous work is introduced. Experimental learning results using a fast camera are presented. Some simple parts (i.e. circular, squared and radiused-square) were used for comparing different connectionist models (Backpropagation, Perceptron and FuzzyARTMAP) and to select the appropriate model. Later during experiments, complex figures were learned using the chosen FuzzyARTMAP algorithm showing a 93.8% overall efficiency and 100% recognition rate. Recognition times were lower than 1 ms, which clearly indicates the suitability of the approach to be implemented in real-world operations.

Keywords: Artificial neural networks, invariant object recognition, machine vision, robotics.

En celdas de ensamble robotizado en ambientes no estructurados, por ejemplo con localización de partes desconocidas, el robot tiene, no solamente que localizar la parte, sino también reconocerla para su agarre. El objetivo de esta investigación es desarrollar un enfoque rápido y robusto para completar la tarea. El enfoque basado en RNA y un reducido conjunto de patrones recurrentes de entrenamiento que aumentan la tarea de reconocimiento comparado con nuestro trabajo es introducido. Se presentan los resultados de aprendizaje experimental utilizando una cámara rápida. Algunas partes simples (es decir, circulares, cuadrados y semi-cuadrado) fueron utilizados para comparar diferentes modelos conexionistas (Backpropagation, Perceptrón y FuzzyARTMAP) y para seleccionar el modelo apropiado. Más tarde, durante los experimentos, se aprendieron figuras complejas mediante el algoritmo de FuzzyARTMAP elegido mostrando un 93,8% tasa de reconocimiento global de eficiencia y un 100% en la razón de reconocimiento. Los tiempos de reconocimiento fueron inferiores a 1 ms, lo que indica claramente la idoneidad del enfoque para implementarse en operaciones de mundo real.

Palabras clave: Redes neuronales artificiales, reconocimiento invariante de objetos, visión de máquina, robótica.

DESCARGAR ARTÍCULO EN FORMATO PDF

Acknowledgements

The authors wish to thank CONACyT through project research grant No. 61373-Y.

References

1. Peña-Cabrera, M., Lopez-Juarez, I., Rios-Cabrera, R., & Corona-Castuera, J. (2005). Machine vision approach for robotic assembly. Assembly Automation, 25(3), 204–216.         [ Links ]

2. Corona-Castuera, J. & Lopez-Juarez, I. (2006). Behaviour-based approach for skill acquisition during assembly operations, starting from scratch. Robotica, 24(6), 657–671.         [ Links ]

3. Hoska DR. (1988) Fixturless assembly manufacturing. Manufacturing Eng, 100:49-54.         [ Links ]

4. Ngyuen, W. & Mills, J.K. (1996). Multi-robot control for flexible fixtureless assembly of flexible sheet metal autobody parts. IEEE International Conference on Robotics and Automation, San Francisco, California, 3, 2340–2345.         [ Links ]

5. Langley, C.S. & D'Eleuterio, G.M.T. (2003). ART FCMAC: a memory efficient neural network for robotic pose estimation. IEEE International Symposium on Computational Intelligence in Robotics and Automation, Kobe, Japan, 1, 418–423.         [ Links ]

6. Aguado, A.S., Montiel, E., & Nixon, M.S. (2002). Invariant characterization of the Hough Transform for pose estimation of arbitrary shapes. Pattern Recognition, 35(5), 1083–1097.         [ Links ]

7. Chin-Hsiung, W., Shi-Jinn, H., & Pei-Zong, L. (2001). A new computation of shape moments via quadtree decomposition. Pattern Recognition, 34(7), 1319-1330.         [ Links ]

8. Best, P.J. & McKay, N.D. (1992). A Method for Registration of 3-D Shapes. IEEE Transactions on Pattern Analysis and Machine Intelligence, 14(2), 239–256.         [ Links ]

9. Bribiesca, E. (1999). A new Chain Code. Pattern Recognition, 32(2), 235–251.         [ Links ]

10. Gonzalez-Galvan, E.J., Korde, U.A., Chen, W., & Skaar, S.B. (1997). Application of a Precision-Enhancing Measure in 3D Rigid-Body Positioning using Camera-Space Manipulation. International Journal of Robotics Research, 16(2), 240–257.         [ Links ]

11. Bone, G.M. & Capson, D. (2003). Vision-guided fixtureless assembly of automotive components. Robotics and Computer-Integrated Manufacturing 19(1-2), 79–87.         [ Links ]

12. Chen, K. (1990). Efficient parallel algorithms for computation of two-dimensional image moments. Pattern Recognition 23(1-2), 109–119.         [ Links ]

13. Carpenter, G.A., Grossberg, S., Markuzon, N., Reynolds, J.H., & Rosen, D.B. (1992). Fuzzy ARTMAP: A Neural Network Architecture for Incremental Supervised Learning of Analog Multidimensional Maps. IEEE Transactions on Neural Networks, 3(5), 698–713.         [ Links ]

14. Carpenter, G.A. & Grossberg, S. (2002). Adaptive Resonance Theory. The Handbook of Brain theory and neural networks, 2nd ed., (87-89), MIT Press.         [ Links ]