Yifan Sun

Phd student in Robotics of CMU. Working on safe motion planning & control, safe RL. @ICL_at_CMU @CMU_Robotics

[Yaru Niu @yaru_niu]

A touch-aware humanoid manipulation policy that cleans the lab for you🧹🧪

Introducing Humanoid Touch Dream: a real-world system for dexterous, contact-rich humanoid loco-manipulation.

Our key idea is simple: the policy predicts future hand forces and tactile latents alongside actions, within a single-stage training framework.

https://t.co/Pt5pXA65wm

1/7

[Yiyuan Pan @pan_yiyuan]

Can task structure emerge directly from robot demonstrations?

We introduce ENAP: a bi-level neuro-symbolic policy that learns an interpretable automaton from visuomotor trajectories and uses it to guide continuous control.

Instead of treating long-horizon manipulation as a pure black box, ENAP recovers task phases, branching behaviors, and recovery loops directly from data.

Project page: https://t.co/ZTDOiNM97N
Paper: https://t.co/mwpX9FAicn

[Yifan Sun @YifanSun98]

[6/6] 📉 In real-world deployment, the learned monitor tracks the robot’s changing stop risk over time and can support proactive intervention before the robot enters an unrecoverable state.

This turns E-stop from a passive last resort into an online safety monitor for humanoids.

Huge thanks to the CMU team:
@YifanSun98 @pan_yiyuan @ShangtaoLi @ChangliuL
and Siemens collaborators:
Caiwu Ding, Tao Cui, Lingyun Wang🙏

#HumanoidSafety #RobotLearning #DigitalTwin

[Yifan Sun @YifanSun98]

[1/6] 🚨 If we press the emergency-stop button now, would the humanoid survive the stop?

Our new work, Learning Safe-Stoppability Monitors for Humanoid Robots, studies exactly this question.

Unlike fixed-base arms, humanoids cannot simply cut power without risking catastrophic falls. Instead, safe stopping depends on whether the robot’s current state still allows a fallback controller to bring it to a minimum-risk condition (MRC).

🎬 Video: https://t.co/gPWxPT7yjB
🌐 Project page: https://t.co/f8vbBnm7rs
📄 Paper: https://t.co/elrMNwiBnl
@ICL_at_CMU @CMU_Robotics @CarnegieMellon @UnitreeRobotics @SiemensUSA
#Robotics #HumanoidRobots #RobotSafety #Sim2Real

[Yifan Sun @YifanSun98]

[5/6] 🎯 The key challenge is data efficiency.

Most nominal states are trivially stoppable. The rare but safety-critical samples live near the boundary between stoppable and unstoppable states.

So PRISM does not waste simulation budget uniformly. It iteratively focuses new labels near this boundary, improving unsafe-state prediction while reducing hazardous real-world data collection.

[Hongyi Chen @chen_hongyi_]

How far can we push dexterous robot manipulation with human video-only supervision and minimal assumptions?
🚫 No teleop. 🚫 No wearables. 🚫 No external sensors. 🚫 No robot demos.
Introducing VIDEOMANIP: 🎥 Just monocular RGB, 🌍 in-the-wild human video → dexterous robot manipulation 🤚[1/6]

[Yifan Sun @YifanSun98]

[5/5] 🧑‍🏫SPARK v1.0 is used as the homework platform for CMU 16-883: Provably Safe Robotics and is designed to be extensible for research and benchmarking (https://t.co/wShWx2eOEv).

Huge thanks to all contributors who made this release possible 🙏

@YifanSun98
@RuiChen_Rob
@kai_s_yun
@dino_fang271
@sebb_ene_
@Li_liangzhii
@Bowei_Li
@weiyezha
@ChangliuL

[Yifan Sun @YifanSun98]

[1/5] 🚀 SPARK v1.0 released!

SPARK (Safe Protective and Assistive Robot Kit) is a modular benchmark + toolkit for safety-aware robot control, spanning teleoperation and autonomy across humanoids and manipulators.

🔗 https://t.co/kf3Dd3gUib
🌐https://t.co/1e879Hh9f4
🎬https://t.co/Gi10ugt0Nk 📄 https://t.co/QDauLlMYvU
👇 What’s new in v1.0 ⤵️

@ICL_at_CMU
@CMU_Robotics
@CarnegieMellon
@UnitreeRobotics
#Robotics #HumanoidSafety #AIInnovation

[Yifan Sun @YifanSun98]

[4/5] 🛡️ Policy-level safety integration Safety is now embedded directly into the policy stack (spark_policy.safe), supporting:

SSA, CBF, SSS (optimization-based)
SMA, PFM (non-optimization)

With reusable safety indices across robots and tasks.

[Hongyi Chen @chen_hongyi_]

[1/7] Teaching dexterous robot hands to perform functional grasps usually needs hours of teleoperation, manual labeling, or pre-scanning object meshes.

Not anymore.
🔥We are excited to introduce Web2Grasp that learns functional multi-finger grasps straight from web images of human hand-object interactions (HOI).
No human demos. No object scans. Just web images.
👉https://t.co/lOcWgMOpu5
@CMU_Robotics @CarnegieMellon

[Yifan Sun @YifanSun98]

Is there any way to apply linear controller to extremely nonlinear robots like humanoid? Come and check out our latest work of Incrementally Koopman approach!🤩

[Feihan Li @Li_liangzhii]

🤖Can highly nonlinear legged robots be controlled by simple linear controller?
🌟Yes—with Incremental Koopman!

Our Incremental Koopman approach transforms nonlinear dynamics into linear models, enabling powerful, proven control techniques. By iteratively expanding datasets and latent spaces, it ensures precise convergence to true system dynamics with monotonic error reduction.

✅This project delivers scalable, high-performance control across terrains and legged robots like Unitree & ANYmal.
🦾A solution for linearized control in locomotion!

Please see our website for the paper and more details:
🌐Website: https://t.co/vUH78l1nZ7

@ICL_at_CMU @CMU_Robotics @CarnegieMellon

#Robotics #Humanoid #AIInnovation

[Unitree @UnitreeRobotics]

Unitree B2-W Talent Awakening! 🥳
One year after mass production kicked off, Unitree’s B2-W Industrial Wheel has been upgraded with more exciting capabilities.
Please always use robots safely and friendly.
#Unitree #Quadruped #Robotdog #Parkour #EmbodiedAI #IndustrialRobot #InspectionRobot #IntelligentRobot #FoundationModels #LeggedRobot #WheeledLegs

[Zhou Xian @zhou_xian_]

Everything you love about generative models — now powered by real physics!

Announcing the Genesis project — after a 24-month large-scale research collaboration involving over 20 research labs — a generative physics engine able to generate 4D dynamical worlds powered by a physics simulation platform designed for general-purpose robotics and physical AI applications.

Genesis's physics engine is developed in pure Python, while being 10-80x faster than existing GPU-accelerated stacks like Isaac Gym and MJX. It delivers a simulation speed ~430,000 faster than in real-time, and takes only 26 seconds to train a robotic locomotion policy transferrable to the real world on a single RTX4090 (see tutorial: https://t.co/bEkIlCKqdf).

The Genesis physics engine and simulation platform is fully open source at https://t.co/DhBv7NdyqH. We'll gradually roll out access to our generative framework in the near future.

Genesis implements a unified simulation framework all from scratch, integrating a wide spectrum of state-of-the-art physics solvers, allowing simulation of the whole physical world in a virtual realm with the highest realism.

We aim to build a universal data engine that leverages an upper-level generative framework to autonomously create physical worlds, together with various modes of data, including environments, camera motions, robotic task proposals, reward functions, robot policies, character motions, fully interactive 3D scenes, open-world articulated assets, and more, aiming towards fully automated data generation for robotics, physical AI and other applications.

Open Source Code: https://t.co/DhBv7NdyqH
Project webpage: https://t.co/SBNyhFB0yn

Documentation: https://t.co/3yuBoaealV

1/n

[Yifan Sun @YifanSun98]

@ElijahGalahad @RuiChen_Rob @kai_s_yun @dino_fang271 @sebb_ene_ @weiyezha @ChangliuL 😀We applied over 50 collision pairs in the experiments. Since all the constraints are linear in control, (5) can be solved with QP which does not have local optima. Check the following paper for another example of solving safe control problem with QP: https://t.co/B4NjBvTZjN

[Yifan Sun @YifanSun98]

[1/4] 🌟Sneak Peek: SPARK in Action! 🦾
Previewing Safe Protective & Assistive Robot Kit (SPARK)—a modular toolbox designed to enhance safety in humanoid autonomy and teleoperation.

Safety isn't just a feature—it's the foundation for humanoids to truly integrate into human life. SPARK filters risky actions, ensuring humanoids can achieve their objectives securely across tasks—from lab experiments to real-world deployments. With SPARK, you can innovate fearlessly, knowing safety is always guaranteed.

Powered by Safe Set Algorithm (SSA), SPARK is built to:
✅ Configure safety behaviors with ease
⚖️ Balance safety and performance
🤖 Integrate with Unitree G1 + Apple Vision Pro
🔧 Support customization for other systems

Stay tuned for the full release in a few weeks 🚀

Please see our website for the paper and more details!
🌐 Website: https://t.co/1e879HhH4C

@ICL_at_CMU @CMU_Robotics @CarnegieMellon @UnitreeRobotics

#Robotics #HumanoidSafety #AIInnovation

[Yifan Sun @YifanSun98]

Team: @YifanSun98 @RuiChen_Rob @kai_s_yun @dino_fang271 @sebb_ene_ @weiyezha @ChangliuL

Safe Set Algorithm (SSA): https://t.co/YevY5uXV4X

Synthesizing provably safe controller: https://t.co/XhE6TUqOLz

Safe control of quadruped in varying dynamics:
https://t.co/3hc9vv1ud3

[Yifan Sun @YifanSun98]

[4/4] Safety is also guaranteed when the humanoid and obstacles are all in motion.