aiqbench analysis
This research introduces PAC-MAN, an RL framework that integrates control-barrier functions with realistic onboard sensing to ensure whole-body safety for humanoid robots during high-speed dodgeball tasks. The study demonstrates that fixed-camera depth segmentation alone provides sufficient observability for evasion strategies without requiring privileged state access or complex multi-sensor setups in deployment scenarios. By successfully zero-shotting a lightweight policy on the Unitree G1, the work establishes practical pathways for safe human-robot interaction under imperfect perception conditions.
PAC-MAN: Perception-Aware CBF-RL for Whole-Body Safety in Humanoid Dodgeball
arXiv:2607.28623v1 Announce Type: new Abstract: We present PAC-MAN, a perception-aware CBF-RL framework that couples control-barrier safety with deployment-realistic onboard sensing for whole-body humanoid dodgeball. The deployed policy sees the ball only as segmentation-masked depth from a head-mounted camera, while training-time CBF guidance represents clearance to every body link, and an adversarial motion prior regularizes the resulting evasive reflexes. We evaluate on a controlled any-link contact benchmark with seeded throws in two regimes: single throws and a deployment loop in which the robot walks back to its station and recovers between throws. On this benchmark, the policy comes within a few points of a privileged state oracle: a fixed onboard camera alone is adequate for evasion. We find that usable barrier structure depends on perceptual observability: Joint-CBF gives the best performance with accurate ball states, degrades under fixed-camera observations when used only as training guidance, and recovers with a ball-tracking gimbal or privileged runtime filter. We therefore deploy a lightweight Link-CBF policy zero-shot on the Unitree G1 in the real world, where it tolerates imperfect perception, succeeds on 95% of throws, and uses semantic segmentation to dodge different balls.
Abstract:We present PAC-MAN, a perception-aware CBF-RL framework that couples control-barrier safety with deployment-realistic onboard sensing for whole-body humanoid dodgeball. The deployed policy sees the ball only as segmentation-masked depth from a head-mounted camera, while training-time CBF guidance represents clearance to every body link, and an adversarial motion prior regularizes the resulting evasive reflexes. We evaluate on a controlled any-link contact benchmark with seeded throws in two regimes: single throws and a deployment loop in which the robot walks back to its station and recovers between throws. On this benchmark, the policy comes within a few points of a privileged state oracle: a fixed onboard camera alone is adequate for evasion. We find that usable barrier structure depends on perceptual observability: Joint-CBF gives the best performance with accurate ball states, degrades under fixed-camera observations when used only as training guidance, and recovers with a ball-tracking gimbal or privileged runtime filter. We therefore deploy a lightweight Link-CBF policy zero-shot on the Unitree G1 in the real world, where it tolerates imperfect perception, succeeds on 95% of throws, and uses semantic segmentation to dodge different balls.
Robotics (cs.RO); Artificial Intelligence (cs.AI)
arXiv:2607.28623 [cs.RO]
(or arXiv:2607.28623v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2607.28623 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 279 words
Related coverage
Dyna-2: General Purpose Physical Agents with Hour-Long Autonomy — Jason Ma, Dyna Robotics
Latent Space · Oct 13, 2026
Filter-Aware Fine-Tuning for Safe Humanoid Whole-Body Tracking
arXiv · Oct 5, 2026
PGMT: Perceptive General Motion Tracking for Humanoid Robots
arXiv · Sep 9, 2026
Perceptive Behavior Foundation Model: Adapting Human Motion Priors to Robot-Centric Terrain
arXiv · Jun 16, 2026