A simulated humanoid uses its arms to brace against walls and recover from pushes up to 100 N while walking, surviving 70.6% of tested push scenarios versus 16.8% for a stepping-only controller.
Force-and-moment-based model predictive control for achieving highly dynamic locomotion on bipedal robots,
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
-
Bracing for Impact: Robust Humanoid Push Recovery and Locomotion with Reduced Order Models
A simulated humanoid uses its arms to brace against walls and recover from pushes up to 100 N while walking, surviving 70.6% of tested push scenarios versus 16.8% for a stepping-only controller.