Detail publikace

USING Q-LEARNING FOR FOUR-LEGGED ROBOT WALKING CONTROL

Originální název

USING Q-LEARNING FOR FOUR-LEGGED ROBOT WALKING CONTROL

Anglický název

USING Q-LEARNING FOR FOUR-LEGGED ROBOT WALKING CONTROL

Jazyk

en

Originální abstrakt

Significant goal in walking robot design is realization of autonomous system which is capable of motion in unknown environment. One of the possibilities how to reach necessary adaptability of control system without modeling of complex and unpredictable cases of system behaviour is the application of machine learning. From various learning paradigms the Q-learning is particularly attractive. For efficient and successful use of Q-learning the proper definition and discretization of state space of environment and state space of robot, suitable choice of action set is essential, and also the set of states and actions should be as small as possible.

Anglický abstrakt

Significant goal in walking robot design is realization of autonomous system which is capable of motion in unknown environment. One of the possibilities how to reach necessary adaptability of control system without modeling of complex and unpredictable cases of system behaviour is the application of machine learning. From various learning paradigms the Q-learning is particularly attractive. For efficient and successful use of Q-learning the proper definition and discretization of state space of environment and state space of robot, suitable choice of action set is essential, and also the set of states and actions should be as small as possible.

BibTex


@inproceedings{BUT9743,
  author="Tomáš {Březina} and Pavel {Houška} and Vladislav {Singule}",
  title="USING Q-LEARNING FOR FOUR-LEGGED ROBOT WALKING CONTROL",
  annote="Significant goal in walking robot design is realization of autonomous system which is capable of motion in unknown environment. One of the possibilities how to reach necessary adaptability of control system without modeling of complex and unpredictable cases of system behaviour is the application of machine learning. From various learning paradigms the Q-learning is particularly attractive. For efficient and successful use of Q-learning the proper definition and discretization of state space of environment and state space of robot, suitable choice of action set is essential, and also the set of states and actions should be as small as possible.",
  address="Institute of Mechanics of Solids
Faculty of Mechanical Engineering
Brno University of Technologi",
  booktitle="Mechtronics, Robotics and Biomechanics 2003",
  chapter="9743",
  institution="Institute of Mechanics of Solids
Faculty of Mechanical Engineering
Brno University of Technologi",
  year="2003",
  month="march",
  pages="103",
  publisher="Institute of Mechanics of Solids
Faculty of Mechanical Engineering
Brno University of Technologi",
  type="conference paper"
}