arxiv.org/abs/2308.13988…
A Compact Variable Stiffness Actuator for Agile Legged Locomotion
C's Robotics Paper Notes
512 posts
C's Robotics Paper Notes
@RoboReading
busy with paperwork recently
Joined March 2024
- arxiv.org/pdf/2506.12779 From Experts to a Generalist: Toward General Whole-Body Control for Humanoid Robots training separate experts for different clusters of motions, (also treat sim2real separately), then distill to one WBC.
- RL from Physical Feedback: Aligning Large Motion Models with Humanoid Control beingbeyond.github.io/RLPF/ use physical simulation to verify and finetune text2motion models
- arxiv.org/pdf/2504.17249 berkeley humanoid lite actuator-level accessible humanoid
- arxiv.org/abs/2505.20829 Learning Unified Force and Position Control for Legged Loco-Manipulation The hybrid interface of doing both position tracking and force control for the end effector enables more stable loco-mani especially for teleoperation.
- Improving Vision-Language-Action Model with Online Reinforcement Learning arxiv.org/abs/2501.16664 online RL for VLA is unstable, so the authors do iterative RL here to stabilize it.
- arxiv.org/pdf/2508.21043 HITTER: A HumanoId Table TEnnis Robot via Hierarchical Planning and Learning Model based planning (with mocap perception) + RL stylistic tracking
- AnyDexGrasp: Learning General Dexterous Grasping for Any Hands with Human-level Learning Efficiency graspnet.net/anydexgrasp/ amazing. title explains all. potential contacts as the intermediate layer.
- kungfu bot arxiv.org/pdf/2506.12851 motion filtering + adaptive tracking rewards for kungfu motion tracking
- arxiv.org/pdf/2504.21738 LangWBC: Language-directed Humanoid Whole-Body Control via End-to-end Learning achieved by teacher student
- MolmoAct: Action Reasoning Models that can Reason in Space arxiv.org/pdf/2508.07917 VLM/VLA with manipulation things and visual reasoning capabilities
- Latent Action Diffusion for Cross-Embodiment Manipulation arxiv.org/pdf/2506.14608 shared latent space DP with different en/decoders for different morph
- BeyondMimic: From Motion Tracking to Versatile Humanoid Control via Guided Diffusion arxiv.org/abs/2508.08241 train tracking policies -> distill with diffusion -> use custom scores for downstream tasks without retraining. It's the followup of their previous SIGGRAPH work at RAI
- Narrate2Nav: Real-Time Visual Navigation with Implicit Language Reasoning in Human-Centric Environments arxiv.org/abs/2506.14233 teacher VLA (with language narration of scene as input) + student distillation to implicitly learn the cues with real-time inference.
