Carregant...
Carregant...

Vés al contingut (premeu Retorn)

Handling high parameter dimensionality in reinforcement learning with dynamic motor primitives

Autor
Colomé, A.; Alenyà, G.; Torras, C.
Tipus d'activitat
Presentació treball a congrés
Nom de l'edició
2013 ICRA Workshop on Novel Methods for Learning and Optimization of Control Policies and Trajectories for Robotics
Any de l'edició
2013
Data de presentació
2013
Llibre d'actes
Proceedings orf the 2013 ICRA Workshop on Novel Methods for Learning and Optimization of Control Policies and Trajectories for Robotics
Projecte finançador
IntellAct: Intelligent observation and execution of Actions and manipulations
Repositori
http://hdl.handle.net/2117/25142 Obrir en finestra nova
URL
http://www.robot-learning.de/Research/ICRA2013 Obrir en finestra nova
Resum
Dynamic Motor Primitives (DMP) are nowadays widely used as movement parametrization for learning trajectories, because of their linearity in the parameters, rescalation robustness and continuity. However, when learning a movement with DMP, where a set of gaussians distributed along the trajectory is used to approximate an acceleration excitation function, a very large number of gaussian approximations need to be performed. Adding them up for all joints yields too many parameters to be explored, ...
Citació
Colomé, A.; Alenyà, G.; Torras, C. Handling high parameter dimensionality in reinforcement learning with dynamic motor primitives. A: ICRA Workshop on Novel Methods for Learning and Optimization of Control Policies and Trajectories for Robotics. "Proceedings orf the 2013 ICRA Workshop on Novel Methods for Learning and Optimization of Control Policies and Trajectories for Robotics". Karlsruhe: 2013.
Paraules clau
intelligent robots, learning by demonstration, manipulators, robot kinematics, robot programming Author keywords: dynamic motor primitives
Grup de recerca
ROBiri - Grup de Robòtica de l'IRI

Participants

Arxius