Learning Human-Humanoid Coordination for Collaborative Object Carrying

์ €์ž: Yushi Du, Yixuan Li, Baoxiong Jia, Yutang Lin, Pei Zhou, Wei Liang, Yanchao Yang, Siyuan Huang | ๋‚ ์งœ: 2025-10-16 | DOI: 10.48550/arXiv.2510.14293 📄 PDF


Essence

Figure 2

Fig. 2: Overview of COLA. Our Policy mainly consists of three steps: (i) We train a base whole-body control policy to pr

COLA๋Š” proprioception๋งŒ์„ ์‚ฌ์šฉํ•˜๋Š” reinforcement learning ๊ธฐ๋ฐ˜์˜ ์ •์ฑ…์œผ๋กœ, humanoid ๋กœ๋ด‡์ด ์ธ๊ฐ„๊ณผ ํ˜‘๋ ฅํ•˜์—ฌ ๋ฌผ์ฒด๋ฅผ ์šด๋ฐ˜ํ•  ๋•Œ ์ ์‘์ ์ด๊ณ  ์•ˆ์ •์ ์ธ whole-body coordination์„ ๊ฐ€๋Šฅํ•˜๊ฒŒ ํ•œ๋‹ค.

Motivation

Achievement

Figure 1

Fig. 1: COLA provides a proprioception-only policy that enables compliant human-humanoid collaboration for carrying dive

How

Figure 2

Fig. 2: Overview of COLA. Our Policy mainly consists of three steps: (i) We train a base whole-body control policy to pr

Originality

Limitation & Further Study

Evaluation

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

์ดํ‰: COLA๋Š” humanoid-human collaborative carrying์ด๋ผ๋Š” ์‹ค์šฉ์  ๊ณผ์ œ์— ๋Œ€ํ•ด proprioception-only ์ •์ฑ…์œผ๋กœ ์™„์ „ํ•œ ์†”๋ฃจ์…˜์„ ์ œ์‹œํ•˜๋ฉฐ, three-step training framework์™€ implicit force modeling์„ ํ†ตํ•ด ๋†’์€ ๋…์ฐฝ์„ฑ์„ ๋ณด์—ฌ์ค€๋‹ค. ์‹œ๋ฎฌ๋ ˆ์ด์…˜๊ณผ ์‹ค์ œ ํ™˜๊ฒฝ์—์„œ ๋™์‹œ์— ๊ฒ€์ฆ๋œ ๊ฒฐ๊ณผ๋Š” ์‹ค์ œ ๋ฐฐํฌ ๊ฐ€๋Šฅ์„ฑ์„ ์‹œ์‚ฌํ•˜๋ฉฐ, human user study๋ฅผ ํ†ตํ•œ compliant collaboration ํ™•์ธ์œผ๋กœ ์‹ค๋ฌด์  ๊ฐ€์น˜๋ฅผ ์ž…์ฆํ•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˜์ •