Learning Smooth Humanoid Locomotion through Lipschitz-Constrained Policies

์ €์ž: Zixuan Chen, Xialin He, Yen-Jen Wang, Qiayuan Liao, Yanjie Ze, Zhongyu Li, S. Shankar Sastry, Jiajun Wu, Koushil Sreenath, Saurabh Gupta, Xue Bin Peng | ๋‚ ์งœ: 2024-10-15 | URL: https://arxiv.org/abs/2410.11825 📄 PDF


Essence

Figure 2

Fig. 2: Lipschitz continuity is a method of quantifying the

๋ณธ ๋…ผ๋ฌธ์€ Reinforcement Learning์œผ๋กœ ํ›ˆ๋ จํ•œ humanoid robot์˜ locomotion policy์— Lipschitz ์ œ์•ฝ์„ ๋ถ€์—ฌํ•˜์—ฌ smooth behavior๋ฅผ ์ž๋™์œผ๋กœ ์œ ๋„ํ•˜๋Š” Lipschitz-Constrained Policies (LCP) ๋ฐฉ๋ฒ•์„ ์ œ์•ˆํ•œ๋‹ค.

Motivation

Achievement

Figure 1

Fig. 1: Lipschitz-constrained policies (LCP) provide a simple and general method for training policies to produce smooth

How

Figure 3

Fig. 3: Gradient of policies trained with and without smooth-

Originality

Limitation & Further Study

Evaluation

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

์ดํ‰: Lipschitz constraint์„ ํ†ตํ•œ smooth policy ํ•™์Šต์€ ์ด๋ก ์ ์œผ๋กœ ๋ช…ํ™•ํ•˜๊ณ  ์‹ค์šฉ์ ์ด๋ฉฐ, ๊ธฐ์กด์˜ ๋ณต์žกํ•œ smoothing ๊ธฐ๋ฒ•๋“ค์„ ๋‹จ์ˆœํ•˜๊ณ  ๋ฏธ๋ถ„ ๊ฐ€๋Šฅํ•œ ๋ฐฉ์‹์œผ๋กœ ๋Œ€์ฒดํ•˜๋Š” ์šฐ์ˆ˜ํ•œ ๊ธฐ์—ฌ์ด๋‹ค. ์‹ค์ œ humanoid robot์—์„œ์˜ ๊ฒ€์ฆ๊ณผ ์žฌํ˜„์„ฑ ์žˆ๋Š” ๊ณต๊ฐœ ์ฝ”๋“œ ๊ณต๊ฐœ๋กœ high impact์„ ๊ธฐ๋Œ€ํ•  ์ˆ˜ ์žˆ๋‹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˜์ •