Koushil Sreenath

Safe-Stop enables humanoid robots to make safer emergency stop decisions by evaluating stoppability in real-time, rather than relying on fixed maneuvers.

ODEWorld achieves high-quality long-horizon predictions and planning-oriented dynamics by seamlessly integrating continuous-time modeling with ODE-based representations.

Synthesizing 48,000 interaction trajectories without human input enables a humanoid robot to learn complex loco-manipulation tasks effectively.

Rigid body trajectory optimization can now avoid singularities and achieve faster convergence with a new structure-aware method that respects the underlying geometry.

LadderMan enables humanoid robots to climb ladders and manipulate objects with unprecedented robustness and adaptability in real-world scenarios.

Steering worst-case trajectories with an adversarial network and Boltzmann reweighting dynamics ensembles yields a surprisingly stable and efficient approach to robust RL under dynamics uncertainty.

Humanoid robots can now perform vision-based parkour, chaining together dynamic skills like climbing, vaulting, and rolling, adapting to real-time obstacle changes.