Pinned Tweet
Humanoid parkour on unseen, extremely challenging terrains... even replicating Boston Dynamics-style moves? 🤖
Introducing TTT-Parkour: A Real-to-Sim-to-Real framework enabling robots to master stakes, beams, and wedges in under 10 minutes! ⏱️
🌐 Project: ttt-parkour.github.io
Shaoting Zhu retweeted
What should control policies and VLAs look like in the era of agentic robotics?
Policies (e.g., VLAs and WAMs) should know their capability limits, call an agent for help, and turn every intervention into greater autonomy.
Introducing Equilibrium Energy Control (EqC): control with introspection, built for self-improving robots. 🤖
🧠 One energy landscape. Both action and introspection.
Using an energy-based action generator, EqC generates actions while monitoring its own competence through the geometry of action refinement. The policy itself knows when to ask for help.
🔄 Energy-Gated DAgger × Astra: an agent-powered self-improvement loop.
Detect a competence gap → request agent recovery → learn from it → expand autonomy. A concrete recipe for RSI in VLAs. More effective than human judged intervention.
• Real-world garment folding: 0% → 100% autonomous success after just 6 Astra-generated recoveries.
• Real-world block stacking: 0% → 96.7%.
• Real-world debris cleaning: 100% success.
🚀 Stronger control across four benchmarks.
EqC boosts π₀ on RoboTwin 2.0 Hard (18% → 33%) and LIBERO-Plus (53.6% → 65.8%) in a plug-and-play manner!
Accepted as a Spotlight at the Continually Self-Improving Robots (CSIR) Workshop @ CoRL 2026!
Come see the poster—and join for my spotlight talk:
🗓 Nov 12, 9:25–9:30 AM CT
📍 Room 202, JW Marriott Austin
Code is open source!
📄 Paper: arxiv.org/abs/2610.18790
🌐 Project & demos: equilibrium-energy-control.g…
💻 GitHub: github.com/hrh6666/EqC
#RobotLearning #AI #LLM #Agents #RSI #corl2026 #corl #VLA #Control #Robotics #EBM
🤖Can whole-body pre-training help humanoids move and manipulate?
Introducing WB-WAM: a World Action Model trained on 1,880 hours of heterogeneous data, with a three-stage recipe bridging human motion and real-robot tasks.
🌐 Project: wb-wam.github.io
Shaoting Zhu retweeted
🤖✨Excited to share our new work:
OMG: Omni-Modal Motion Generation for Generalist Humanoid Control
What if a humanoid could understand intent from language, music/audio, human motion, or their combinations—and turn it into executable whole-body motion in real time? [🧵1/11]
Shaoting Zhu retweeted
Flexible Locomotion Learning with Diffusion Model Predictive Control
Excited to share that our paper has been accepted to #ICRA2026 @ieee_ras_icra!
A diffusion-planning framework for flexible real-world quadruped locomotion. Instead of learning a fixed RL policy or relying on hand-crafted dynamics for MPC, we train a diffusion trajectory prior that jointly predicts future states and actions.
Key Ideas:
Diffusion-MPC: A diffusion planner unlocks flexible locomotion through test-time reward and constraint adaptation
Interactive reward-weighted finetuning enables continual behavior refinement from online environment feedback
Real-world deployment on Unitree Go2 with efficient and adaptive planning
The same planner can adapt at test time to height changes, posture/joint constraints, balancing under external disturbances, energy-aware locomotion, and zero-shot outdoor walking on grass and slopes.
🌐Homepage: flexible-diffusion-mpc.githu…
📖Paper: arxiv.org/abs/2510.04234
🔗Code: github.com/hrh6666/Flexible-…
This work is by @RunhanH, Haldun Balim, @hankyang94 , and @du_yilun.
#ICRA2026 #Robotics #LeggedRobots #RobotLearning #DiffusionModels #MPC #MachineLearning
See our full-stack, instinct-level whole-body control toolkit for legged & humanoid robots at:
nitter.cf/ziwenzhuang_leo/status…
We believe robots need instinct, not only reasoning.
Introducing Project-Instinct — a full-stack, instinct-level whole-body control toolkit for legged & humanoid robots.
🔗 project-instinct.github.io
(1/3)
🚀 Hiking in the Wild: A Scalable Perceptive Parkour Framework for Humanoids!
Excited to release our latest work! We push the boundaries of humanoid agility in unstructured environments.
Highlights:
• Zero-Shot Sim-to-Real: No mapping, no external state estimation.
• Edge-Aware Safety: Novel volumetric edge penalization prevents slipping on terrain edges.
• Extremely dynamic: Robust traversal at up to 2.5 m/s!
• Open-sourced: All codes for training, deployment are open-sourced!
👇 Check out the project page!
project-instinct.github.io/h…
Shaoting Zhu retweeted
🚀 VR-Robo: A Real-to-Sim-to-Real pipeline for RGB vision-based navigation & control in legged robots.
💡 Reconstruct realistic indoor scenes using RGB
🧠 Train RL policies with photorealistic simulation
🤖 Deploy directly on real visual robots!
🔗 vr-robo.github.io
🧠+🐶 = Terrain Navigation
Introducing SARO: a Space-Aware Robot System that combines Vision-Language Models and RL-based control for robust quadruped navigation in 3D terrains 🌄
Catch us at Thursday 15:15pm at Room 309
👉 saro-vlm.github.io/
🚨 New work at #ICRA2025!
Robust Robot Walker 🐾
We enable quadruped robots to pass tiny traps (bars, pits, poles) using only proprioception – no cameras, no depth!
Catch us at Thursday 16:55pm in Room 305!
🔗 robust-robot-walker.github.i…
③/④
During deployment, our policy achieves approximate omnidirectional movement by joystick commands without motion capture or other auxiliary localization techniques. A joystick is used to generate fake goal commands, consisting of constant values for Δ𝐺 and Δ𝑡.
1
102
②/④
We use explicit-implicit dual-state learning. The contact force is first encoded by a contact encoder to an implicit latent, and concatenated with explicit privileged state to the dual-state. In addition, we introduce a classification head to guide the policy in learning.
1
119
Shaoting Zhu retweeted
Enjoy generalizable robot segmentation and plug-and-play visual augmentation🥳!
RoboEngine enables visual generalization and robustness in almost all scenes with data collected from just one scene! The simplest idea, but useful and effective😂.
roboengine.github.io/









