Mostra el registre d'ítem simple

dc.contributorAbbeel, Pieter
dc.contributorTorras, Carme
dc.contributor.authorClavera Gilaberte, Ignasi
dc.contributor.otherInstitut de Robòtica i Informàtica Industrial
dc.date.accessioned2017-07-07T11:40:45Z
dc.date.available2017-07-07T11:40:45Z
dc.date.issued2017-05
dc.identifier.urihttp://hdl.handle.net/2117/106257
dc.description.abstractNon-prehensile manipulation, such as pushing, is an important function for robots to move objects and is sometimes preferred as an alternative to grasping. However, due to unknown frictional forces, pushing has been proven a difficult task for robots. We explore the use of reinforcement learning to train a robot to robustly push an object. In order to deal with the sample complexity of training such a method, we train the pushing policy in simulation and then transfer this policy to the real world. In order to ease the transfer from simulation, we propose to use modularity to separate the learned policy from the raw inputs and outputs; rather than training ``end-to-end," we decompose our system into modules and train only a subset of these modules in simulation. We further demonstrate that we can incorporate prior knowledge about the task into the state space and the reward function to speed up convergence. Finally, we introduce "reward guiding" to modify the reward function and further reduce the training time. We demonstrate, in both simulation and real-world experiments, that such an approach can be used to reliably push an object from many initial positions and orientations.
dc.language.isoeng
dc.publisherUniversitat Politècnica de Catalunya
dc.rights.urihttp://creativecommons.org/licenses/by-nc-nd/3.0/es/
dc.subjectÀrees temàtiques de la UPC::Informàtica::Intel·ligència artificial
dc.subjectÀrees temàtiques de la UPC::Matemàtiques i estadística
dc.subject.lcshArtificial intelligence
dc.subject.otherDeep Reinforcement Learning
dc.subject.otherDeep Learning
dc.subject.otherArtificial Intelligence
dc.subject.otherReinforcement Learning
dc.subject.otherRobotics
dc.subject.otherMachine Learning
dc.titlePolicy transfer via modularity
dc.typeBachelor thesis
dc.subject.lemacIntel·ligència artificial
dc.subject.amsClassificació AMS::68 Computer science::68T Artificial intelligence
dc.identifier.slugFME-1375
dc.rights.accessOpen Access
dc.date.updated2017-07-06T21:11:23Z
dc.audience.educationlevelGrau
dc.audience.mediatorUniversitat Politècnica de Catalunya. Facultat de Matemàtiques i Estadística
dc.audience.degreeGRAU EN MATEMÀTIQUES (Pla 2009)
dc.contributor.covenanteeUniversity of California Berkeley. Electrical Engineering and Computer Science
dc.description.mobilityOutgoing


Fitxers d'aquest items

Thumbnail

Aquest ítem apareix a les col·leccions següents

Mostra el registre d'ítem simple