DIAL: Distilling Intent-Aware Latents for Vision-Language-Action on Humanoid Robots

์ €์ž: | ๋‚ ์งœ: 2026-03-31 | URL: https://arxiv.org/list/cs.RO/current 📄 PDF


Essence

Figure 2

Fig. 2.

SoftHand Model-W๋Š” 3D ํ”„๋ฆฐํŒ… ๊ธฐ๋ฐ˜์˜ ์ธ๊ฐ„ํ˜• ๋กœ๋ด‡ ์†์œผ๋กœ, 2-DoF ์†๋ชฉ์„ ํ†ตํ•ฉํ•˜์—ฌ ์†๊ฐ€๋ฝ์˜ underactuated tendon-driven ๊ตฌ์กฐ์™€ ์†๋ชฉ์˜ ๋Šฅ๋™์  ์ œ์–ด๋ฅผ ๊ฒฐํ•ฉํ–ˆ๋‹ค. Carpal tunnel ์˜๊ฐ์˜ ํž˜์ค„ ๋ผ์šฐํŒ…์„ ํ†ตํ•ด ์›๊ฒฉ ๋ชจํ„ฐ ๋ฐฐ์น˜๋ฅผ ๊ฐ€๋Šฅํ•˜๊ฒŒ ํ•˜๋ฉด์„œ compactํ•œ ํ˜•ํƒœ๋ฅผ ์œ ์ง€ํ•œ๋‹ค.

Motivation

Achievement

Figure 1

Fig. 1.

How

Figure 3

Fig. 3.

Originality

Limitation & Further Study

Evaluation

Novelty: 4/5 Technical Soundness: 4/5 Significance: 4/5 Clarity: 4/5 Overall: 4/5

์ดํ‰: SoftHand Model-W๋Š” soft robotics์˜ adaptive synergies ๊ฐœ๋…์„ ์œ ์ง€ํ•˜๋ฉด์„œ ๋Šฅ๋™์  ์†๋ชฉ์„ ์ฒ˜์Œ ํ†ตํ•ฉํ•œ ํ˜์‹ ์  ์„ค๊ณ„์ด๋ฉฐ, 3D ํ”„๋ฆฐํŒ…๊ณผ carpal tunnel routing์„ ํ†ตํ•ด ์‹ค์šฉ์„ฑ๊ณผ anthropomorphism์„ ๋™์‹œ์— ๋‹ฌ์„ฑํ–ˆ๋‹ค. ์†๋ชฉ ์ถ”๊ฐ€์˜ ๋ช…ํ™•ํ•œ ์„ฑ๋Šฅ ๊ฐœ์„  ํšจ๊ณผ๋ฅผ ์ž…์ฆํ•˜์—ฌ dexterous manipulation ๋ถ„์•ผ์— ์˜๋ฏธ ์žˆ๋Š” ๊ธฐ์—ฌ๋ฅผ ํ•œ๋‹ค.

← ๋ชฉ๋ก์œผ๋กœ ๋Œ์•„๊ฐ€๊ธฐ

๐ŸŽง Audio Overview

์ด ๋…ผ๋ฌธ ๋ฆฌ๋ทฐ๋ฅผ ํŒŸ์บ์ŠคํŠธํ˜• ์˜ค๋””์˜ค๋กœ ์ƒ์„ฑํ•ฉ๋‹ˆ๋‹ค. (Gemini ยท ํ‚ค๋Š” ๋ธŒ๋ผ์šฐ์ €์—๋งŒ ์ €์žฅ ยท ์™„์„ฑ๋ณธ์€ ์ด๋ฉ”์ผ๋กœ๋„ ์ „์†ก)
โ–ธ ๊ณ ๊ธ‰: ๊ตฌ์„ฑ ๋ฐฉํ–ฅ(๋Œ€๋ณธ ์ž‘์„ฑ ์ง€์นจ) ์ง์ ‘ ์ˆ˜์ •