
这届 IROS 2026 上,具身学术与产业边界正在消融
Quick Answer
IROS 2026 highlights a shift in embodied AI towards integrating physical laws into models, with key trends including the adoption of 'big and small brain' architectures and tactile feedback as core to dexterous manipulation.
Quick Take
Innovations like the ULTRA framework and LITHE architecture demonstrate significant advancements in real-time control and data efficiency, making embodied robotics more accessible.
Key Points
- ULTRA framework integrates physical-driven neural algorithms for full-body humanoid control.
- LITHE architecture enables dynamic switching between and real-time control with under 100 microseconds latency.
- Tactile feedback is now essential for dexterous manipulation, surpassing visual limitations.
- Sim2Real approaches reduce data collection costs significantly, enabling broader access for smaller teams.
- Chinese teams are increasingly influential, showcasing innovative solutions in embodied intelligence.
DeepSignal Analysis
What happened
IROS 2026 showcased significant advancements in embodied AI, emphasizing the integration of physical laws into models. Key trends included the adoption of 'big and small brain' architectures and the central role of tactile feedback in dexterous manipulation. Innovations like the ULTRA framework and LITHE architecture improved real-time control and data efficiency.
Key evidence
- The workshop on Physical World Models for Scaling Embodied AI highlighted the need for models to internalize physical laws, moving beyond mere semantic understanding.
- The LITHE paper demonstrated a layered architecture where Python handles high-level tasks while C++ manages real-time control, achieving execution times under 100 microseconds.
- The HALO framework achieved over 80% transfer rate from simulation to reality for heavy-load humanoid robots, significantly reducing data collection costs.
Why it matters
The shift towards integrating physical principles into AI models marks a critical evolution in robotics, moving from theoretical concepts to practical applications. This focus on real-world interactions and efficiency could democratize access to advanced robotics technology, enabling smaller teams to compete with larger firms. The emphasis on tactile feedback also suggests a growing recognition of its importance in complex manipulation tasks.
📖 Reader Mode
~4 min read
作者丨齐铖湧
编辑丨岑 峰

01
具身基模:别再 “语义拟合” 了,得懂物理







02
“大小脑” 架构:
今年终于成了机器人顶会标配




03
触觉,终于坐上顶会正桌了




04
顶会不仅在关注数据采集,
并且找到了 “新平替”

05
总结:
工程落地,已经成为顶会新风向


为了方便大家抱团交流、互通会议消息,我们专门建了 IROS 2026 参会交流群!进群你会解锁这些「福利」?
会议情报站:投稿 DDL、格式规范、评审进度……关键节点第一时间送达,再也不怕错过 deadline,投稿准备稳稳的。
同行茶话会:天南海北的同行都在群里,聊心得、碰思路、攒人脉,科研路上我们同行。
前沿补给站:最新研究成果、热点方向实时同步,结合会议主题帮你捋清投稿脉络,给论文灵感加满油。
现场后援团:(会议期间专属开启):到了会场也不用慌——
路线导航:场馆分布、报告厅怎么走,群里随时问;
议题共读:热门 session、Keynote 探讨,错过也能追;
Poster 汇编:现场海报精华整理,一键收藏不遗漏;
约局广场:约饭局、采访局、交流局……想唠嗑、想合作、想面基,群里发起就成。
? 进群传送门: 扫码进群或添加微信Luyoyo_2026,备注:ECCV + 单位/学校+姓名+方向。

搞科研/搞技术,信息差很重要。
来,一起快人一步!

上车,带你看遍全球 AI 顶会精华
可独家畅览:
专家演讲PPT
大会报告全文
热门论文解读
学术新星访谈

扫描上方二维码
或点击「阅读原文」关注专区。
雷峰网原创文章,未经授权禁止转载。详情见转载须知。
— Originally published at leiphone.com
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from 雷峰网 AI
See more →
刚刚,GPT 5.6 发布会上,OpenAI 暴露了哪些 Agent 技术路线?
OpenAI's GPT 5.6 integrates ChatGPT and Codex, introducing a for complex task execution, with models Soul, Terra, and Luna for efficient workflow management. The release emphasizes task orchestration, contextual understanding, and robust security measures for enterprise applications.

