On-line knowledge acquisition and enhancement in robotic assembly tasks

Industrial robots are reliable machines for manufacturing tasks such as welding, panting, assembly, palletizing or kitting operations. They are traditionally programmed by an operator using a teach pendant in a point-to-point scheme with limited sensing capabilities such as industrial vision systems and force/torque sensing. The use of these sensing capabilities is associated to the particular robot controller, operative systems and programming language. Today, robots can react to environment changes specific to their task domain but are still unable to learn skills to effectively use their current knowledge. The need for such a skill in unstructured environments where knowledge can be acquired and enhanced is desirable so that robots can effectively interact in multimodal real-world scenarios.In this article we present a multimodal assembly controller (MAC) approach to embed and effectively enhance knowledge into industrial robots working in multimodal manufacturing scenarios such as assembly during kitting operations with varying shapes and tolerances. During learning, the robot uses its vision and force capabilities resembling a human operator carrying out the same operation. The approach consists of using a MAC based on the Fuzzy ARTMAP artificial neural network in conjunction with a knowledge base. The robot starts the operation having limited initial knowledge about what task it has to accomplish. During the operation, the robot learns the skill for recognising assembly parts and how to assemble them. The skill acquisition is evaluated by counting the steps to complete the assembly, length of the followed assembly path and compliant behaviour. The performance improves with time so that the robot becomes an expert demonstrated by the assembly of a kit with different part geometries. The kit is unknown by the robot at the beginning of the operation; therefore, the kit type, location and orientation are unknown as well as the parts to be assembled since they are randomly fed by a conveyor belt. A novel multimodal assembly controller (MAC) was designed having minimal assembly information (-Z assembly direction).It was demonstrated that knowledge can be refined when other different matting pair are assembled without the need to acquire another primitive knowledge base (PKB).The robot learns a new assembly and improves its skills from experience observed by a reduced number of patterns, lower compliant forces and shorter assembly trajectories.The MAC demonstrated that can be used in non-structured environments for the kitting process with uncertainties (in position and geometry) for both, the kits and the pegs.

[1]  Angel P. del Pobil,et al.  Sensor-based learning for practical planning of fine motions in robotics , 2002, Inf. Sci..

[2]  Hendrik Van Brussel,et al.  Compliant Robot Motion I. A Formalism for Specifying Compliant Motion Tasks , 1988, Int. J. Robotics Res..

[3]  Ismael Lopez-Juarez,et al.  Machine vision approach for robotic assembly , 2005 .

[4]  Louis A. Martin-Vega,et al.  Industrial perspective on research needs and opportunities in manufacturing assembly , 1995 .

[5]  Jude W. Shavlik,et al.  Knowledge-Based Artificial Neural Networks , 1994, Artif. Intell..

[6]  Rüdiger Dillmann,et al.  Experience-based optimization of universal manipulation strategies for industrial assembly tasks , 2011, Robotics Auton. Syst..

[7]  Dima Damen,et al.  Egocentric Real-time Workspace Monitoring using an RGB-D camera , 2012, 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems.

[8]  Vasant Honavar,et al.  Symbolic Artificial Intelligence and Numeric Artificial Neural Networks: Towards A Resolution of the Dichotomy , 1995 .

[9]  J. Y. Kim.  Novel visual sensing systems for overcoming occlusion in robotic assembly , 2005 .

[10]  Masatoshi Ishikawa,et al.  Fast peg-and-hole alignment using visual compliance , 2013, 2013 IEEE/RSJ International Conference on Intelligent Robots and Systems.

[11]  Ulrike Thomas,et al.  Multi Sensor Fusion in Robot Assembly Using Particle Filters , 2007, Proceedings 2007 IEEE International Conference on Robotics and Automation.

[12]  JakovljevicZivana,et al.  Fuzzy inference mechanism for recognition of contact states in intelligent robotic assembly , 2014 .

[13]  Ismael Lopez-Juarez,et al.  On the design of intelligent robotic agents for assembly , 2005, Inf. Sci..

[14]  Stephen Grossberg,et al.  Fuzzy ARTMAP: A neural network architecture for incremental supervised learning of analog multidimensional maps , 1992, IEEE Trans. Neural Networks.

[15]  V. Gullapalli,et al.  Acquiring robot skills via reinforcement learning , 1994, IEEE Control Systems.

[16]  Rüdiger Dillmann,et al.  Building elementary robot skills from human demonstration , 1996, Proceedings of IEEE International Conference on Robotics and Automation.

[17]  Stephen Grossberg,et al.  ARTMAP: supervised real-time learning and classification of nonstationary data by a self-organizing neural network , 1991, [1991 Proceedings] IEEE Conference on Neural Networks for Ocean Engineering.

[18]  E. Sadauskas,et al.  ALIGNMENT OF THE PARTS USING HIGH FREQUENCY VIBRATIONS , 2013 .

[19]  Stephen Grossberg,et al.  A massively parallel architecture for a self-organizing neural pattern recognition machine , 1988, Comput. Vis. Graph. Image Process..

[20]  Sébastien Gérard,et al.  Towards a core ontology for robotics and automation , 2013, Robotics Auton. Syst..

[21]  Ales Ude,et al.  Analysis of human peg-in-hole executions in a robotic embodiment using uncertain grasps , 2013, 9th International Workshop on Robot Motion and Control.

[22]  Emilio Miguelanez,et al.  An IEEE standard Ontology for Robotics and Automation , 2012, 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems.

[23]  Angel P. del Pobil,et al.  Robot Physical Interaction through the combination of Vision, Tactile and Force Feedback - Applications to Assistive Robotics , 2012, Springer Tracts in Advanced Robotics.

[24]  Torsten Kröger,et al.  Parallel Stiffness Actuators with Six Degrees of Freedom for Efficient Force/Torque Control Applications , 2011, Robotic Systems for Handling and Assembly.