PIMBS adds physics-informed losses to body schema learning for musculoskeletal robots, reducing the amount of training data needed, with clear gains in simulation and smaller, tuned gains on a real humanoid.
ECCE1: the first of a series of anthropomimetic musculoskeletal upper torsos,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.RO 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
PIMBS: Efficient Body Schema Learning for Musculoskeletal Humanoids with Physics-Informed Neural Networks
PIMBS adds physics-informed losses to body schema learning for musculoskeletal robots, reducing the amount of training data needed, with clear gains in simulation and smaller, tuned gains on a real humanoid.