상세 보기
A Novel Approach for Efficient Gaussian Mixture Model using Dynamics-motivated Optimal Excitation
- Kim, Taehoon;
- Jeong, Juwon;
- Kong, Taejune;
- Lee, Hyunwook;
- Oh, Sehoon
Citations
WEB OF SCIENCE
0Citations
SCOPUS
0초록
EffiDynaMix, a novel, efficient, gray-box, nonparametric dynamics modeling method, integrates mathematical dynamics with Gaussian Mixture Model (GMM) for simplified training data creation and improved generalization. It outperforms traditional methods like conv-GMM, GP, and LSTM in training efficiency and accuracy with new data. By leveraging dynamic equations, EffiDynaMix enhances learning efficiency and adaptability, offering advancements in robotic system precision and computational efficiency, leading to faster and more responsive robots. © 2024 IEEE.
키워드
IDENTIFICATION; LIKELIHOOD; TASK
- 제목
- A Novel Approach for Efficient Gaussian Mixture Model using Dynamics-motivated Optimal Excitation
- 저자
- Kim, Taehoon; Jeong, Juwon; Kong, Taejune; Lee, Hyunwook; Oh, Sehoon
- 발행일
- 2024-07
- 유형
- Proceedings Paper
- 저널명
- IEEE International Symposium on Industrial Electronics