Tianhao Wei

PhD at Carnegie Mellon. @ICL_at_CMU @CMU_Robotics @CMU_ECE. Robotics, safe control, embodied AI, neural network verification.

[Tianhao Wei @wei_tianhao]

🤔 How can a robot accomplish heterogeneous tasks, with precision and guarantees?

🚀 Introducing Meta-Control: Automating model-based control synthesis for heterogeneous skills with LLM!

https://t.co/zeN9aoiqG5

[Xiaofeng Guo @Xiaofeng2Guo]

🚀Can we have a freely moving hand in the air for the manipulation policy to directly command in the real world?

We introduce Flying Hand: End-Effector-Centric Framework for Versatile Aerial Manipulation Teleoperation and Policy Learning.

🎯EE-centric MPC for aerial manipulator
🤝EE-centric teleoperation and imitation learning for aerial manipulation
✅Various common tasks achieved

🤖We hope to bring aerial manipulation back to the view of the general manipulation community.

Website: https://t.co/NogYRYXdqI

[Tairan He @TairanHe99]

🚀 Can we make a humanoid move like Cristiano Ronaldo, LeBron James and Kobe Byrant?

YES!

🤖 Introducing ASAP: Aligning Simulation and Real-World Physics for Learning Agile Humanoid Whole-Body Skills

Website: https://t.co/XQga7tIfdw
Code: https://t.co/NpEeJtVxpp

[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

[Tianhao Wei @wei_tianhao]

@sabeerawa05 @ICL_at_CMU @CMU_Robotics @SCSatCMU @CMUEngineering @CarnegieMellon @Tsinghua_Uni Thank you!

[Tianhao Wei @wei_tianhao]

🤔 How can a robot accomplish heterogeneous tasks, with precision and guarantees?

🚀 Introducing Meta-Control: Automating model-based control synthesis for heterogeneous skills with LLM!

https://t.co/zeN9aoiqG5

[Tianhao Wei @wei_tianhao]

Compositional system is the key for versatile and reliable control synthesis. We share a similar insight in our recent work Meta-Control. Congrats @ShuoCheng94 @AjayMandlekar !

[Shuo Cheng @ShuoCheng94]

Can we teach a robot hundreds of tasks with only dozens of demos?
Introducing NOD-TAMP: A framework that chains together manipulation skills from as few as one demo per skill to compositionally generalize across long-horizon tasks with unseen objects and scenes. (1/N)

[Tianhao Wei @wei_tianhao]

Opening cabinets can be unexpectedly hard for robots. It's one of the major motivations of our recent work on LLM empowered model-based cobtrol. Congrats @arjun__gupta !

[Arjun Gupta @arjun__gupta]

Introducing: Opening Cabinets and Drawers in the Real World using a Commodity Mobile Manipulator

We develop a system to open unseen cabinets and drawers *zero-shot* from novel environments using the Stretch RE2: https://t.co/ZueIGMOBvZ

[Michał Zawalski @MiZawalski]

🤖Can robots think through complex tasks step-by-step like language models?
We present Embodied Chain-of-Thought Reasoning (ECoT): enabling robots to reason about plans and actions for better performance🎯, interpretability🧐, and generalization🌎.

See https://t.co/jRkwpRZswZ.

[Tianhao Wei @wei_tianhao]

@ChongZitaZhang Haha, very good point

[Tianhao Wei @wei_tianhao]

Zero-shot deployment withou human crafted skills is very important. Congrats @DJiafei !

[Jiafei Duan @DJiafei]

🚀 Excited to share our latest work: MANIPULATE-ANYTHING! 🦾 This scalable method pushes the boundaries of real-world robotic manipulation through zero-shot task execution and automated BC data generation. Here's a quick overview:👇

https://t.co/TrrY5iyy6H

[Tianhao Wei @wei_tianhao]

Language encodes so many things about how the world works that we are even not aware of. There is definitely much more to explore.

[Sergey Levine @svlevine]

Chain-of-thought reasoning is a powerful tool to enable language models to work through complex problems. Can we use this with robots? With embodied chain-of-thought, vision-language-action (VLA) models can think through perception and planning!

A 🧵👇

[Tianhao Wei @wei_tianhao]

Very impressive work. Congrats @Ying_yyyyyyyy @QinYuzhe !

[Ying Yuan @Ying_yyyyyyyy]

Code released for Robot Synesthesia: 
In-Hand Manipulation with Visuotactile Sensing!
Project page: https://t.co/kebcxWfmVD
Code: https://t.co/I1OJUHTlqv

[Tianhao Wei @wei_tianhao]

Glad to see robots are actually making the world better. Congrats @priyasun_ !

[Priya Sundaresan @priyasun_]

Introducing FLAIR: Feeding via Long-horizon AcquIsition of Realistic dishes! Our system merges a library of skills with foundation models for efficient robotic feeding, tailored to user preferences.

🔗https://t.co/3Y6ZnG7u00
📃https://t.co/NQseBY1gsN
To Appear at RSS ’24

1/N

[Tianhao Wei @wei_tianhao]

@TairanHe99 Thank you, Tairan!

[Xiaolong Wang @xiaolonw]

Imagine if you have an object in hand, you can rotate the object by feeling without even looking.

This is what we enable the robot to do now: Rotating without Seeing. Our multi-finger robot hand learns to rotate diverse objects using only touch sensing.

https://t.co/1jJE3DbytT

[Tianhao Wei @wei_tianhao]

Audio for tactile sensing is defintely underexplored. I still remember how amazed I was when I first learned this direction from prof. @OliverKroemer. Congrats @Liu_Zeyi_ !

[Zeyi Liu @Liu_Zeyi_]

🔊 Audio signals contain rich information about daily interactions. Can our robots learn from videos with sound?

Introducing ManiWAV, a robotic system that learns contact-rich manipulation skills from in-the-wild audio-visual data. See thread for more details (1/4) 👇

[Tianhao Wei @wei_tianhao]

Very interesting work, congrats @Xiaofeng2Guo !

[Xiaofeng Guo @Xiaofeng2Guo]

We introduce 𝐅𝐥𝐲𝐢𝐧𝐠 𝐂𝐚𝐥𝐥𝐢𝐠𝐫𝐚𝐩𝐡𝐞𝐫, an aerial manipulation system that can draw various calligraphy artworks:

🎯Contact-aware trajectory planning and hybrid control
✏️Intuitive user interface and novel end-effector design
🧑‍🎨UAM can draw letters with changing linewidth

Website: https://t.co/9jZbM5xPHy

[Tianhao Wei @wei_tianhao]

Tactile sensing sim2real is a crucial step in generating Internet-level data for robotics. Congrats @HaozhiQ !

[Haozhi Qi @HaozhiQ]

Introducing tactile skin sim-to-real for dexterous in-hand translation!

We propose a simulation model for ReSkin, a magnetic tactile sensing skin. It can simulate ternary shear and binary normal forces.

More: https://t.co/fp2c86lnOZ

[Tianhao Wei @wei_tianhao]

A big step towards robot foundation model! Congrats @TairanHe99

[Tairan He @TairanHe99]

Introduce OmniH2O, a learning-based system for whole-body humanoid teleop and autonomy:
🦾Robust loco-mani policy
🦸Universal teleop interface: VR, verbal, RGB
🧠Autonomy via @chatgpt4o or imitation
🔗Release the first whole-body humanoid dataset
https://t.co/wzbzyjhAoc

[Tianhao Wei @wei_tianhao]

Meta-Control harnesses LLM's power to automate expert thought processes, creating customized model-based control for heterogeneous skills, paving the way for universal robotic foundations.

[Tianhao Wei @wei_tianhao]

🦾 Diverse heterogeneous manipulation tasks? ✔️
🛠️ Customized state representations & control strategies? ✔️
🤖 Achieving higher autonomy without human intervention? ✔️
📊 Rigorous analysis, generalizability, & robustness? ✔️
⏱️ Real-time, efficient reliable execution? ✔️