Han Zhang

122 posts

Han Zhang banner
Han Zhang

Han Zhang

@DoubleHan07

PhD student in robotics @Stanford | Ham radio (DE BI1NWO 73) | Prev @Tsinghua_Uni

Palo Alto, CA Katılım Ekim 2013
241 Takip Edilen410 Takipçiler
Sabitlenmiş Tweet
Han Zhang
Han Zhang@DoubleHan07·
Collecting manipulation data with DexUMI. Let’s scale up together! 📈 Big shoutout to my amazing project co-lead Mengda @mengdaxu__ , who brought full stack experience, strong system design skills, and unique insight to make this possible! 🙌
Mengda Xu@mengdaxu__

Can we collect robot dexterous hand data directly with human hand? Introducing DexUMI: 0 teleoperation and 0 re-targeting dexterous hand data collection system → autonomously complete precise, long-horizon and contact-rich tasks Project Page: dex-umi.github.io

English
1
3
22
3.1K
Han Zhang retweetledi
Suning Huang
Suning Huang@suning_huang·
🤖Low-data post-training can teach a VLA policy a new robot skill. But it also makes it too attached to the training demos. We call this lock-in🔒: the policy can execute the post-training task, yet fails to respond to seemingly obvious prompt changes. DeLock preserves steerability using only the policy’s own pretrained knowledge. No extra supervision needed!🚀🚀🚀 #Robotics #AI #EmbodiedAI #VLA
English
5
44
177
30.7K
Han Zhang retweetledi
Xiaomeng Xu
Xiaomeng Xu@XiaomengXu11·
Can we learn whole-body mobile manipulation directly from human demonstrations? Introducing Whole-Body Mobile Manipulation Interface (HoMMI) Egocentric + UMI, 0 teleop -> bimanual & whole-body manipulation, long-horizon navigation, active perception hommi-robot.github.io
English
11
72
324
65.2K
Han Zhang retweetledi
Sunday
Sunday@sundayrobotics·
Introducing ACT-1: The frontier robot AI behind Memo. ACT-1 enables Memo to perform: 🎭 ⏰ Lengthy tasks like cleaning up after your dinner 🦾 Dexterous tasks such as folding socks ☕ Daily delights like making an espresso How does it work? 🧵
English
41
101
890
200.9K
Han Zhang retweetledi
Cheng Chi
Cheng Chi@chichengcc·
mm level precision beyond actuator limits, so much torque that you need to manage thermals. Owning the whole stack from HW to AI is the only way 🦾
English
25
23
399
110.1K
Han Zhang retweetledi
Haochen Shi
Haochen Shi@HaochenShi74·
Toddy enjoyed his time in Seoul — his first big trip away from home!😆 He rolled around on wheels, showed off his pull-up skills (Cr: @li_yitang), spoke in the spotlight as a 1.5-year-old, and presented at the poster session.
Haochen Shi tweet mediaHaochen Shi tweet mediaHaochen Shi tweet mediaHaochen Shi tweet media
Haochen Shi@HaochenShi74

ToddlerBot is accepted to CoRL, and we will bring Toddy (2.0 version) to Seoul for a trip. Come and say hi to Toddy if you're around😁! Our arxiv paper is also updated with more technical details in the appendix: arxiv.org/abs/2502.00893

English
3
7
77
12.1K
Han Zhang retweetledi
Hao-Shu Fang
Hao-Shu Fang@haoshu_fang·
How do we unlock the full dexterity of robot hands with data, even beyond what teleoperation can achieve? DEXOP captures natural human manipulation with full-hand tactile & proprio sensing, plus direct force feedback to users, without needing a robot👉dex-op.github.io
English
31
272
1.4K
197.6K
Han Zhang
Han Zhang@DoubleHan07·
@kikori1906 Wow made my night! Thank you Nina and Taka!😊
English
1
0
1
56
Han Zhang
Han Zhang@DoubleHan07·
@ja7kpi We (BY1QH) are on 2m MS now. We will adjust our transceiver settings for your setup.
English
0
0
7
125
Tada/JA7KPI
Tada/JA7KPI@ja7kpi·
ペルセウス座流星群MS、RN0JJ(約1300km)までは届いたようです。依然として Even/1stでCQを出し続けていますが、交信は未だできていません。現在、アンテナは北京方面(西)に向けています。
日本語
1
0
4
403
Han Zhang retweetledi
RoboPapers
RoboPapers@RoboPapers·
Collecting dexterous humanoid robot data is difficult to scale. That's why @mengdaxu__ and @DoubleHan07 built DexUMI: a tool for demonstrating how to control a dexterous robot hand, which allows you to quickly collect task data. Co-hosted by @micoolcho and @chris_j_paxton
English
0
10
23
26.3K
Han Zhang retweetledi
Chris Paxton
Chris Paxton@chris_j_paxton·
Human hands are incredibly capable manipulators, but it's hard to get high-quality robot data for learning policies. DexUMI provides a solution: a tool by which you can naturally "puppet" a dexterous hand with your own, in order to learn skills that directly transfer to a robot. Learn more ->
RoboPapers@RoboPapers

Collecting dexterous humanoid robot data is difficult to scale. That's why @mengdaxu__ and @DoubleHan07 built DexUMI: a tool for demonstrating how to control a dexterous robot hand, which allows you to quickly collect task data. Co-hosted by @micoolcho and @chris_j_paxton

English
5
27
183
18.3K
Han Zhang
Han Zhang@DoubleHan07·
@kikori1906 Sorry to hear that. I’ll be also on air at the following time (all at OL49): Aug.07 13:56z AO-123 15:13z SO-124 Aug.08 00:43z ISS Would these be available then?
English
1
0
1
38
Han Zhang
Han Zhang@DoubleHan07·
@kikori1906 I’m currently at the intersection of both grids. Try to catch me on the air.
English
1
0
1
35
Han Zhang
Han Zhang@DoubleHan07·
@kikori1906 09:42z ISS will be my last operation at OL48 & OL49
English
1
0
0
25
Han Zhang
Han Zhang@DoubleHan07·
@ba1pk Still can not find the proper setting for RS-44. During my operation the downlink signal is very weak. Sorry for that…
English
1
0
0
64
BA1PK
BA1PK@ba1pk·
7.22 BI1NWO OM32, OM33 06:27z ISS 08:02z ISS 09:31z RS-44 145.955 SSB 10:00z PO-101 There is a risk that RS-44 cannot be sent for QSO (the equipment is still under debugging. I'm trying my best).
English
3
0
3
437