Soft actor-critic based design of cochlear implant wireless power transfer system for high efficiency and wide data bandwidth높은 효율과 넓은 데이터 대역폭을 위한 강화학습 기반의 인공와우 무선 전력 전송 시스템 디자인

Cited 0 time in webofscience Cited 0 time in scopus
  • Hit : 54
  • Download : 0
In the WPT system, the coupling coefficient between the Tx and Rx coil is an important design variable. In the case of cochlear implants, however, the coupling coefficient of the WPT system may change due to individual body differences. Therefore, the coupling coefficient-insensitivity is very important design issue at the cochlear implant WPT system. In this paper, I propose a stagger tuning method of a cochlear implant WPT system. The proposed method makes the cochlear implant WPT system stable at the link gain regardless of the coupling coefficient. In addition, the design method can be optimized with consideration of the power transfer efficiency (PTE) of the WPT system. I verified the proposed method using the reinforcement learning based optimal design model. Through the proposed Soft Actor Critic (SAC) based RL model, the cochlear implant WPT system was optimized in the dynamic design environment. In addition, it was confirmed that the proposed SAC algorithm shows significantly better performance compared to other optimization methods like the Proximal Policy Optimization (PPO) and Bayesian Optimization (BO).
Advisors
Kim, Jounghoresearcher김정호researcher
Description
한국과학기술원 :전기및전자공학부,
Publisher
한국과학기술원
Issue Date
2022
Identifier
325007
Language
eng
Description

학위논문(박사) - 한국과학기술원 : 전기및전자공학부, 2022.8,[iii, 44 p. :]

Keywords

Cochlear implant▼aWireless power transfer▼aArtificial intelligence▼aReinforcement learning▼aSystem efficiency▼aBandwidth; 인공와우▼a무선 전력 전송▼a인공지능▼a강화학습▼a시스템 효율▼a대역폭

URI
http://hdl.handle.net/10203/309078
Link
http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=1007868&flag=dissertation
Appears in Collection
EE-Theses_Ph.D.(박사논문)
Files in This Item
There are no files associated with this item.

qr_code

  • mendeley

    citeulike


rss_1.0 rss_2.0 atom_1.0