SafeHumanoid: VLM-RAG-driven Control of Upper Body Impedance for Humanoid Robot

์ €์ž: Yara Mahmoud, Jeffrin Sam, Nguyen Khang, Marcelino Fernando, Issatay Tokmurziyev, Miguel Altamirano Cabrera, Muhammad Haris Khan, Artem Lykov, Dzmitry Tsetserukou | ๋‚ ์งœ: 2025-11-28 | URL: https://arxiv.org/abs/2511.23300 📄 PDF


Essence

Figure 1

Figure 1: Egocentric perception and semantic-to-safety

SafeHumanoid๋Š” Vision Language Model(VLM)๊ณผ Retrieval-Augmented Generation(RAG)์„ ํ™œ์šฉํ•˜์—ฌ ํœด๋จธ๋…ธ์ด๋“œ ๋กœ๋ด‡์˜ ์ž„ํ”ผ๋˜์Šค์™€ ์†๋„๋ฅผ ๋™์ ์œผ๋กœ ์กฐ์ •ํ•˜๋Š” ์‹œ์Šคํ…œ์œผ๋กœ, ์ธ๊ฐ„-๋กœ๋ด‡ ์ƒํ˜ธ์ž‘์šฉ ์‹œ ์•ˆ์ „์„ฑ๊ณผ ์ž‘์—… ์™„๋ฃŒ๋ฅผ ๋™์‹œ์— ๋‹ฌ์„ฑํ•œ๋‹ค.

Motivation

Achievement

Figure 2

Figure 2: SafeHumanoid pipeline architecture. The onboard PC streams egocentric frames and executes impedance control at

How

Figure 2

Figure 2: SafeHumanoid pipeline architecture. The onboard PC streams egocentric frames and executes impedance control at

Originality

Limitation & Further Study

Evaluation

Novelty: 4/5 Technical Soundness: 3/5 Significance: 4/5 Clarity: 4/5 Overall: 4/5

์ดํ‰: SafeHumanoid๋Š” ์˜๋ฏธ๋ก ์  ์ถ”๋ก ๊ณผ ์ž„ํ”ผ๋˜์Šค ์ œ์–ด์˜ ํ˜์‹ ์  ๊ฒฐํ•ฉ์œผ๋กœ ์ธ๊ฐ„-๋กœ๋ด‡ ํ˜‘๋ ฅ์˜ ์•ˆ์ „์„ฑ์„ ํฌ๊ฒŒ ํ–ฅ์ƒ์‹œํ‚ค๋Š” ์ œ์•ˆ์ด์ง€๋งŒ, ์ถ”๋ก  ์ง€์—ฐ์‹œ๊ฐ„๊ณผ ์‹ค์‹œ๊ฐ„์„ฑ์€ ์‹ค์ œ ๋ฐฐํฌ๋ฅผ ์œ„ํ•ด ํ•ด๊ฒฐํ•ด์•ผ ํ•  ์ฃผ์š” ๊ณผ์ œ์ด๋‹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˜์ •