Active Tactile Object Recognition by Monte Carlo Tree Search

This paper considers the problem of object recognition using only tactile information. The focus is on designing a sequence of robot hand grasps that achieves accurate recognition after few enclosure contacts. It seeks to maximize the recognition probability and minimize the number of touches required. The actions are formulated as hand poses relative to each other, making the algorithm independent of small object movements and absolute workspace coordinates. The optimal sequence of actions is approximated by Monte Carlo tree search. We demonstrate active tactile recognition results in physics simulation and on a real robot. In simulation, most object instances were recognized within 16 moves. On a real robot, our method correctly recognized objects in 2--9 grasps and outperformed a greedy baseline.

[1]  Sharon A. Stansfield,et al.  A Robotic Perceptual System Utilizing Passive Vision and Active Touch , 1988, Int. J. Robotics Res..

[2]  Carmel Domshlak,et al.  Simple Regret Optimization in Online Planning for Markov Decision Processes , 2012, J. Artif. Intell. Res..

[3]  J. Schneiter.  An objective tactile sensing strategy for object recognition and localization , 1986, Proceedings. 1986 IEEE International Conference on Robotics and Automation.

[4]  Gaurav S. Sukhatme,et al.  Towards Interactive Object Recognition , 2014 .

[5]  Gregory D. Hager,et al.  Tactile-Object Recognition From Appearance Information , 2011, IEEE Transactions on Robotics.

[6]  David M. Siegel.  Finding the pose of an object in a hand , 1991, Proceedings. 1991 IEEE International Conference on Robotics and Automation.

[7]  W. Grimson,et al.  Model-Based Recognition and Localization from Sparse Range or Tactile Data , 1984 .

[8]  Ruzena Bajcsy,et al.  Active and exploratory perception , 1992, CVGIP Image Underst..

[9]  Leslie Pack Kaelbling,et al.  Task-Driven Tactile Exploration , 2010, Robotics: Science and Systems.

[10]  R. Klatzky,et al.  Hand movements: A window into haptic object recognition , 1987, Cognitive Psychology.

[11]  Sethu Vijayakumar,et al.  Active estimation of object dynamics parameters with tactile sensors , 2010, 2010 IEEE/RSJ International Conference on Intelligent Robots and Systems.

[12]  Andrew W. Moore,et al.  Reinforcement Learning: A Survey , 1996, J. Artif. Intell. Res..

[13]  Andreas Krause,et al.  Near-Optimal Bayesian Active Learning with Noisy Observations , 2010, NIPS.

[14]  Ruzena Bajcsy,et al.  Object Recognition Using Vision and Touch , 1985, IJCAI.

[15]  Peter K. Allen,et al.  Acquisition and interpretation of 3-D sensor data from touch , 1990, IEEE Trans. Robotics Autom..

[16]  Simon M. Lucas,et al.  A Survey of Monte Carlo Tree Search Methods , 2012, IEEE Transactions on Computational Intelligence and AI in Games.

[17]  Joel W. Burdick,et al.  The next best touch for model-based localization , 2013, 2013 IEEE International Conference on Robotics and Automation.

[18]  Nicholas R. J. Lawrance,et al.  Nonmyopic planning for long-term information gathering with an aerial glider , 2014, 2014 IEEE International Conference on Robotics and Automation (ICRA).

[19]  Leslie Pack Kaelbling,et al.  Grasping POMDPs , 2007, Proceedings 2007 IEEE International Conference on Robotics and Automation.

[20]  Tae-Kyun Kim,et al.  Recovering 6D Object Pose and Predicting Next-Best-View in the Crowd , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[21]  M. Ani Hsieh,et al.  A triangle histogram for object classification by tactile sensing , 2016, 2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).

[22]  Kris K. Hauser,et al.  Randomized Belief-Space Replanning in Partially-Observable Continuous Spaces , 2010, WAFR.

[23]  Honglak Lee,et al.  Deep Learning for Real-Time Atari Game Play Using Offline Monte-Carlo Tree Search Planning , 2014, NIPS.

[24]  Andreas Krause,et al.  Adaptive Submodularity: Theory and Applications in Active Learning and Stochastic Optimization , 2010, J. Artif. Intell. Res..

[25]  Miles C. Bowman,et al.  Control strategies in object manipulation tasks , 2006, Current Opinion in Neurobiology.

[26]  Ruzena Bajcsy,et al.  Object exploration in one and two fingered robots , 1987, Proceedings. 1987 IEEE International Conference on Robotics and Automation.

[27]  Oliver Brock,et al.  Interactive Perception: Leveraging Action in Perception and Perception in Action , 2016, IEEE Transactions on Robotics.

[28]  Peter K. Allen,et al.  Integrating Vision and Touch for Object Recognition Tasks , 1988, Int. J. Robotics Res..

[29]  Siddhartha S. Srinivasa,et al.  Efficient touch based localization through submodularity , 2012, 2013 IEEE International Conference on Robotics and Automation.

[30]  Tomás Lozano-Pérez,et al.  Tactile Recognition and Localization Using Object Models: The Case of Polyhedra on a Plane , 1984, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[31]  Kazuo Tanie,et al.  Development of a finger-shaped tactile sensor and its evaluation by active touch , 1992, Proceedings 1992 IEEE International Conference on Robotics and Automation.

[32]  Joel Veness,et al.  Monte-Carlo Planning in Large POMDPs , 2010, NIPS.

[33]  Wolfram Burgard,et al.  Object identification with tactile sensors using bag-of-features , 2009, 2009 IEEE/RSJ International Conference on Intelligent Robots and Systems.

[34]  Dimitri P. Bertsekas,et al.  Approximate Dynamic Programming , 2017, Encyclopedia of Machine Learning and Data Mining.

[35]  Csaba Szepesvári,et al.  Bandit Based Monte-Carlo Planning , 2006, ECML.