Incremental Learning of Skills in a Task-Parameterized Gaussian Mixture Model

详细信息    查看全文

• 作者：Jose Hoyos ; Flavio Prieto ; Guillem Alenyà…
• 关键词：Programming by demonstration ; Robot learning ; Incremental learning
• 刊名：Journal of Intelligent and Robotic Systems
• 出版年：2016
• 出版时间：April 2016
• 年：2016
• 卷：82
• 期：1
• 页码：81-99
• 全文大小：2,512 KB
• 参考文献：1.Alizadeh, T., Calinon, S., Caldwell, D.G.: Learning from demonstrations with partially observable task parameters. In: Proceedings of IEEE International Conference on Robotics and Automation (ICRA), pp. 3309–3314 (2014)
  2.Argall, B.D., Chernova, S., Veloso, M., Browning, B.: A survey of robot learning from demonstration. Rob. Autom. Syst. 57, 469–483 (2009)CrossRef
  3.Billard, A., Calinon, S., Dillmann, R., Schaal, S.: Robot programming by demonstration. In: Bruno Siciliano, O.K. (ed.) Handbook of Robotics, chap. 59, pp 1371–1394. Springer, Berlin (2008)
  4.Breazeal, C., Scassellati, B.: Challenges in building robots that imitate people. In: Dautenhahn, K., Nehaniv, C.L. (eds.) Imitation in Animals and Artifacts, pp. 363–390. MIT Press (2002)
  5.Calinon, S., Alizadeh, T., Caldwell, D.: On improving the extrapolation capability of task-parameterized movement models. In: IEEE International Conference on Intelligent Robots and Systems, pp. 610–616 (2013)
  6.Calinon, S., Billard, A.: Incremental learning of gestures by imitation in a humanoid robot. In: IEEE Human Robot Interation Conference, pp. 255–262 (2007)
  7.Calinon, S., Li, Z., Alizadeh, T., Tsagarakis, N.G., Caldwell, D.: Statistical dynamical systems for skills acquisition in humanoids. In: 8th IEEE-RAS International Conference on Humanoid Robots, pp. 323–329 (2012)
  8.Cederborg, T., Li, M., Baranes, A., Oudeyer, P.Y.: Incremental local online gaussian mixture regression for imitation learning of multiple tasks. In: IEEE International Conference on Intelligent Robots and Systems, pp. 267–274 (2010)
  9.Grollman, D., Jenkins, O.: Dogged learning for robots. In: IEEE International Conference on Robotics and Automation, pp. 2483–2488 (2007)
  10.Hall, P., Hicks, Y.: A method to add gaussian mixture models, Technical Report, University of Bath, p. 18 (2004)
  11.Kulic, D., Takano, W., Nakamura, Y.: Incremental learning, clustering and hierarchy formation of whole body motion patterns using adaptive hidden markov chains. The International Journal of Robotics Research 27(7), 761–784 (2008)CrossRef
  12.Lee, D., Ott, C.: Incremental motion primitive learning by physical coaching using impedance control. In: IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp. 4133–4140 (2010)
  13.Lee, D., Ott, C.: Incremental kinesthetic teaching of motion primitives using the motion refinement tube. In: Autonomous Robot, vol. 31, pp. 115–131. Springer (2011)
  14.Lee, S.H., Kim, H.K., Suh, I.H.: Incremental learning of primitive skills from demonstration of a task. In: 6th ACM/IEEE International Conference on Human-Robot Interaction (HRI), pp. 185–186 (2011)
  15.Niekum, S., Chitta, S., Marthi, B., Osentoski, S., Barto, A.G.: Incremental semantically grounded learning from demonstration. In: Robotics: Science and Systems (2013)
  16.Pardowitz, M., Knoop, S., Dillmann, R., Zollner, R.: Incremental learning of tasks from user demonstrations, past experiences, and vocal comments. IEEE Trans. Syst. Man Cybern. Part B Cybern. 37(2), 322–332 (2007)CrossRef
  17.Rozo, L., Calinon, S., Caldwell, D.: Learning force and position constraints in human-robot cooperative transportation. In: Robot and Human Interactive Communication, 2014 RO-MAN: The 23rd IEEE International Symposium on, pp. 619–624 (2014)
  18.Rozo, L., Calinon, S., Caldwell, D.G., Jimenez, P., Torras, C.: Learning collaborative impedance-based robot behaviors. In: Conference on Artificial Intelligence (AAAI), pp. 1422–1428 (2013)
  19.Townsend, W.T., Guertin, J.A.: Teleoperator slave - WAM design methodology. Industrial Robot: An International Journal 26(3), 167–177 (1999)CrossRef
  
• 作者单位：Jose Hoyos (1)
  Flavio Prieto (2)
  Guillem Alenyà (3)
  Carme Torras (3)
  
  1. Universidad del Quindío, Av Bolivar cl 12N, Armenia, Colombia
  2. Universidad Nacional de Colombia, Carrera 30 No 45-03, Bogota, Colombia
  3. Institut de Robòtica i Informàtica Industrial CSIC-UPC Parc Tecnològic de Barcelona, C/ Llorens i Artigas 4-6, Barcelona, Spain
  
• 刊物类别：Engineering
• 刊物主题：Automation and Robotics
  Electronic and Computer Engineering
  Artificial Intelligence and Robotics
  Mechanical Engineering
  
• 出版者：Springer Netherlands
• ISSN：1573-0409

文摘

Programming by demonstration techniques facilitate the programming of robots. Some of them allow the generalization of tasks through parameters, although they require new training when trajectories different from the ones used to estimate the model need to be added. One of the ways to re-train a robot is by incremental learning, which supplies additional information of the task and does not require teaching the whole task again. The present study proposes three techniques to add trajectories to a previously estimated task-parameterized Gaussian mixture model. The first technique estimates a new model by accumulating the new trajectory and the set of trajectories generated using the previous model. The second technique permits adding to the parameters of the existent model those obtained for the new trajectories. The third one updates the model parameters by running a modified version of the Expectation-Maximization algorithm, with the information of the new trajectories. The techniques were evaluated in a simulated task and a real one, and they showed better performance than that of the existent model.