Shiho Kim
연세대학교 · Computer Science
Shiho Kim 교수의 연구실은 저전력·에너지 자립형 무선 통신 및 인공지능 기반 스마트 시스템의 설계와 최적화를 핵심으로 합니다. RFID 기반의 반도체 칩 설계, 에너지 수확 기술, 열전 발전기용 최적 출력 추적 회로 등 에너지 효율성과 신뢰성을 극대화하는 하드웨어 기반 기술 개발에 집중합니다. 또한, 딥 뉴럴 네트워크를 활용한 자율 주행 제어 알고리즘 개발을 통해 실생활 주행 환경에서의 정밀한 자동 주차 제어를 실현하고자 합니다. 이는 전력 효율성과 지능형 제어 기술의 융합을 통해 미래 스마트 시스템의 핵심 기반 기술을 선도하고자 하는 연구 비전을 담고 있습니다.
표시된 성과는 수집된 데이터 기준으로 산출되며, 일부 차이가 있을 수 있습니다.
A fully integrated passive and battery powered semi-active UHF RFID transponder chip supporting EPC Gen 2 protocol is presented. The proposed transponder works as a passive RFID tag when the generated RF-power is sufficient to operate, otherwise it operates in semi-active mode using battery power. The chip has re-writeable non-volatile memory bank formed by FeRAM and on-chip temperature sensor. The memory consists of EPC memory bank for EPC functionality and temperature bank for storing sensed d
A power source combined with energy harvesting can provide wireless devices for low maintenance cost or extended battery life. We reviewed RF energy harvesting circuits for delivering power to wireless system operating at very low power levels with high efficiency. We need to apply low power control techniques to perform smart management for the wirelessly powered devices.
A digital coreless maximum power point tracking (MPPT) circuit for thermoelectric generator unit is proposed and fabricated. The experimental and simulation results from the proposed MPPT circuit dealt with rapid variation of temperature and abrupt changes of load current have shown that the proposed method allows stable operation with high power transfer efficiency. The proposed MPPT circuit has a merit in cost and miniaturization of a system compared to conventional MPPT algorithms thanks to a
We propose an artificial deep neural network‐ (ANN‐) based automatic parking controller that overcomes a stubborn restriction prevalent in traditional approaches. The proposed ANN learns human‐like control laws for automatic parking through supervised learning from a training database generated by computer‐aided optimizations or real experiments. By learning the relationships between the instantaneous vehicle states and the corresponding maneuver parameters, the proposed twin controller yields l
