Self-organization of Dynamic Object Features Based on Bidirectional Training

This paper presents a method to self-organize object features that describe object dynamics using bidirectional training. The model is composed of a dynamics learning module and a feature extraction module. Recurrent Neural Network with Parametric Bias (RNNPB) is utilized for the dynamics learning module, learning and self-organizing the sequences of robot and object motions. A hierarchical neural network is linked to the input of RNNPB as the feature extraction module for self-organizing object features that describe the object motions. The two modules are simultaneously trained through bidirectional training using image and motion sequences acquired from the robot's active sensing with objects. Experiments are performed with the robot's pushing motion with a variety of objects to generate sliding, falling over, bouncing and rolling motions. The results have shown that the model is capable of self-organizing object dynamics based on the self-organized features.

[1]  Mitsuo Kawato,et al.  A tennis serve and upswing learning robot based on bi-directional theory , 1998, Neural Networks.

[2]  J. Gibson The Ecological Approach to Visual Perception , 1979 .

[3]  Yasuharu Koike,et al.  PII: S0893-6080(96)00043-3 , 1997 .

[4]  Tetsuya Ogata,et al.  Predicting Object Dynamics from Visual Images through Active Sensing Experiences , 2007, Proceedings 2007 IEEE International Conference on Robotics and Automation.

[5]  J. L. Buessler,et al.  Modular neural architectures for robotics , 2003 .

[6]  Maya Cakmak,et al.  The learning and use of traversability affordance using range images on a mobile robot , 2007, Proceedings 2007 IEEE International Conference on Robotics and Automation.

[7]  Jun Tani,et al.  Self-organization of behavioral primitives as multiple attractor dynamics: A robot experiment , 2003, IEEE Trans. Syst. Man Cybern. Part A.

[8]  Tetsuya Ogata,et al.  Object dynamics prediction and motion generation based on reliable predictability , 2008, 2008 IEEE International Conference on Robotics and Automation.

[9]  Alexander Stoytchev,et al.  Behavior-Grounded Representation of Tool Affordances , 2005, Proceedings of the 2005 IEEE International Conference on Robotics and Automation.

[10]  Patrice Wira,et al.  Modular Learning Schemes for Visual Robot Control , 2005, Biomimetic Neural Learning for Intelligent Robots.

[11]  Giorgio Metta,et al.  Grounding vision through experimental manipulation , 2003, Philosophical Transactions of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences.

[12]  James M. Rehg,et al.  Traversability classification using unsupervised on-line visual learning for outdoor robot navigation , 2006, Proceedings 2006 IEEE International Conference on Robotics and Automation, 2006. ICRA 2006..

[13]  Nicholas J. Butko,et al.  Active perception , 2010 .

[14]  Tetsuo Ono,et al.  Robovie: an interactive humanoid robot , 2001 .

[15]  J. Pettigrew,et al.  The effect of visual experience on the development of stimulus specificity by kitten cortical neurones , 1974, The Journal of physiology.

[16]  Manuel Lopes,et al.  Learning Object Affordances: From Sensory--Motor Coordination to Imitation , 2008, IEEE Transactions on Robotics.

[17]  Wenzhi Liao,et al.  Image Feature Extraction Based on Kernel ICA , 2008, 2008 Congress on Image and Signal Processing.

[18]  Yen-Wei Chen,et al.  Image feature representation by the subspace of nonlinear PCA , 2002, Object recognition supported by user interaction for service robots.

[19]  Giorgio Metta,et al.  Early integration of vision and manipulation , 2003, Proceedings of the International Joint Conference on Neural Networks, 2003..

[20]  Tetsuya Ogata,et al.  Experience Based Imitation Using RNNPB , 2006, 2006 IEEE/RSJ International Conference on Intelligent Robots and Systems.

[21]  James L. McClelland,et al.  Parallel distributed processing: explorations in the microstructure of cognition, vol. 1: foundations , 1986 .

[22]  Hans Knutsson,et al.  Canonical correlation analysis in early vision processing , 2001, ESANN.

[23]  Jean-Luc Buessler,et al.  Neurobiology suggests the design of modular architectures for neural control , 2000, Proceedings. 2000 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2000) (Cat. No.00CH37113).

[24]  Michael I. Jordan Attractor dynamics and parallelism in a connectionist sequential machine , 1990 .