Learning Locomotion Skills Using DeepRL: Does the Choice of Action Space Matter?

Simple item page
dc.contributor.authorPeng, Xue Binen_US
dc.contributor.authorPanne, Michiel van deen_US
dc.contributor.editorBernhard Thomaszewski and KangKang Yin and Rahul Narainen_US
dc.date.accessioned2017-12-31T10:45:17Z
dc.date.available2017-12-31T10:45:17Z
dc.date.issued2017
dc.description.abstractThe use of deep reinforcement learning allows for high-dimensional state descriptors, but little is known about how the choice of action representation impacts learning and the resulting performance. We compare the impact of four di erent action parameterizations (torques, muscle-activations, target joint angles, and target jointangle velocities) in terms of learning time, policy robustness, motion quality, and policy query rates. Our results are evaluated on a gaitcycle imitation task for multiple planar articulated figures and multiple gaits. We demonstrate that the local feedback provided by higher-level action parameterizations can signi cantly impact the learning, robustness, and motion quality of the resulting policies.en_US
dc.description.sectionheadersPapers IV: Physics-Based Characters
dc.description.seriesinformationEurographics/ ACM SIGGRAPH Symposium on Computer Animation
dc.identifier.doi10.1145/3099564.3099567
dc.identifier.isbn978-1-4503-5091-4
dc.identifier.issn1727-5288
dc.identifier.pagesXue Bin Peng and Michiel van de Panne-Computing methodologies Animation; Physical simulation; Control methods; Reinforcement learning; physics-based character animation, motion control, locomotion skills
dc.identifier.urihttps://doi.org/10.1145/3099564.3099567
dc.identifier.urihttps://diglib.eg.org:443/handle/10.1145/3099564-3099567
dc.publisherACMen_US
dc.subjectComputing methodologies Animation
dc.subjectPhysical simulation
dc.subjectControl methods
dc.subjectReinforcement learning
dc.subjectphysics
dc.subjectbased character animation
dc.subjectmotion control
dc.subjectlocomotion skills
dc.titleLearning Locomotion Skills Using DeepRL: Does the Choice of Action Space Matter?en_US
Files
Collections
SCA 17: Eurographics/SIGGRAPH Symposium on Computer Animation