Planning, learning, and executing in autonomous systems

  • Ramón García-Martínez
  • Daniel Borrajo
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1348)


Systems that act autonomously in the environment have to be able to integrate three basic behaviors: planning, execution, and learning. Planning involves describing a set of actions that will allow the autonomous system to achieve high utility (a similar concept to goals in high-level classical planning) in an unknown world. Execution deals with the interaction with the environment by application of planned actions and observation of resulting perceptions. Learning is needed to predict the responses of the environment to the system actions, thus guiding the system to achieve its goals. In this context, most of the learning systems applied to problem solving have been used to learn control knowledge for guiding the search for a plan, but very few systems have focused on the acquisition of planning operator descriptions. In this paper, we present an integrated system that learns operator definitions, plans using those operators, and executes the plans for modifying the acquired operators. The results clearly show that the integrated planning, learning, and executing system outperforms the basic planner in a robot domain.


Planning unsupervised machine learning autonomous intelligent systems theory formation and revision 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    M. Barbehenn and S. Hutchinson. An integrated architecture for learning and planning in robotic domains. Sigart Bulletin, 2(4):29–33, 1991.Google Scholar
  2. 2.
    Calistri-Yeh. Classifying and Detecting Plan Based Misconceptions for Robust Plan Recognition. PhD thesis, Department of Computer Science. Brown University, 1990.Google Scholar
  3. 3.
    Jaime G. Carbonell and Yolanda Gil. Learning by experimentation: The operator refinement method. In R. S. Michalski and Y. Kodratoff, editors, Machine Learning: An Artificial Intelligence Approach, Volume III, pages 191–213. Morgan Kaufmann, Palo Alto, CA, 1990.Google Scholar
  4. 4.
    Allan Christiansen. Automatic Acquisition of Task Theories for Robotic Manipulation. PhD thesis, School of Computer Science, Carnegie Mellon University, Pittsburgh, PA, March 1992.Google Scholar
  5. 5.
    B. Falkenhainer. A unified approach to explanation and theory formation. In J. Shrager and Langley P., editors, Computational Models of Scientific Discovery and Theory Formation. Morgan Kaufmann, 1990.Google Scholar
  6. 6.
    W. Fritz, R. García-Martínez, J. Blanqué, A. Rama, R. Adobbati, and M. Samo. The autonomous intelligent system. Robotics and Autonomous Systems, 5(2):109–125, 1989.Google Scholar
  7. 7.
    Ramón García-Martínez. Heuristic theory formation as a machine learning method. In Proceedings of the VI International Symposium on Artificial Intelligence, pages 294–298, México, 1993. LIMUSA.Google Scholar
  8. 8.
    Ramón García-Martínez and Daniel Borrajo. Unsupervised machine learning embedded in autonomous intelligent systems. In Proceedings of the 14th LASTED International Conference on Applied Informatics, pages 71–73, Innsbruck, Austria, 1996.Google Scholar
  9. 9.
    Frederick Hayes-Roth. Using proofs and refutations to learn from experience. In R. S. Michalski, J. G. Carbonell, and T. M. Mitchell, editors, Machine Learning, An Artificial Intelligence Approach, pages 221–240. Tioga Press, Palo Alto, CA, 1983.Google Scholar
  10. 10.
    S. Mahavedan and J. Connell. Automatic programming of behavior-based robots using reinforcement learning. Artificial Intelligence, 55:311–365, 1992.Google Scholar
  11. 11.
    Stephen Salzberg. Heuristics for inductive learning. In Proceedings of the Ninth International Joint Conference on Artificial Intelligence, pages 603–609, Los Angeles, CA, 1985.Google Scholar
  12. 12.
    W. Shen. Discovery as autonomous learning from enviroment. Machine Learning, 12:143–165, 1993.Google Scholar
  13. 13.
    Reid Simmons and Tom M. Mitchell. A task control architecture for mobile robots. In Working Notes of the AAAI Spring Symposium on Robot Navigation., 1989.Google Scholar
  14. 14.
    Peter Stone and Manuela M. Veloso. Towards collaborative and adversarial learning: A case study in robotic soccer. To appear in International Journal of Human-Computer Systems (IJHCS), 1996.Google Scholar
  15. 15.
    Richard Sutton. Integrated architectures for learning, planning, and reacting based on approximating dynamic programming. In Proceedings of the Seventh International Conference on Machine Learning, pages 216–224, Austin, TX, 1990. Morgan Kaufmann.Google Scholar
  16. 16.
    Manuela Veloso, Jaime Carbonell, Alicia Pérez, Daniel Borrajo, Eugene Fink, and Jim Blythe. Integrating planning and learning: The PRODIGY architecture. Journal of Experimental and Theoretical AI, 7:81–120, 1995.Google Scholar
  17. 17.
    Xuemei Wang. Planning while learning operators. In B. Drabble, editor, Proceedings of the Third International Conference on Artificial Intelligence Planning Systems (AIPS96), pages 229–236, Edinburgh, Scotland, May 1996.Google Scholar
  18. 18.
    C. J. C. H. Watkins and P. Dayan. Technical note: Q-learning. Machine Learning, 8(3/4):279–292, May 1992.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1997

Authors and Affiliations

  • Ramón García-Martínez
    • 1
  • Daniel Borrajo
    • 2
  1. 1.Departmento de Computación Facultad de IngenierfaUniversidad de Buenos Aires Bynon 1605. Adrogue (1846)Buenos AiresArgentina
  2. 2.Departamento de Informática Escuela Politécnica SuperiorUniversidad Carlos III de MadridLeganés (Madrid)España

Personalised recommendations