http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
LQ제어기를 이용한 3자유도 시스템 잔류진동의 능동제어
장석범,최진희 외1명 陸軍士官學校 1998 한국군사학논집 Vol.54 No.-
An active control system is designed for a three degree of freedom system using the LQ(Linear Quadratic) design methodology and linear observer. The objective of this study is to design a linear observer which estimates unmeasured state variables, and to design a LQ control system which control the residual vibration of a three degree of freedom system using the observer. Two actuators - the Base Actuator and the Forward Actuator - are considered in this study and compared in performance. The Base Actuator is proved of having fast settling time and small vibration amplitude than the Forward Actuator. The Forward Actuator is proved of having slow settling time and a large vibration amplitude than the Base Actuator. Both actuators are proved of having fast settling time than the condition of LQ control system is not considered. The linear observer estimates the state variables very well in both the Base Actuator and the Forward Actuator. In this study, it is concluded that the LQ control system using the linear observer could be used in controlling of residual vibration in three degree of freedom system and the Base actuator is show better performance than the Forward Actuator in settling time and vibration amplitude.
박성우(Seong-Woo Park),장석범(Seok-beom Jang),오주희(Joo-Hee Oh) 한국정보기술학회 2024 Proceedings of KIIT Conference Vol.2024 No.5
본 연구는 자폐 스펙트럼 장애(ASD)의 조기 진단을 위한 컴퓨터 비전 및 머신러닝 기술의 적용 가능성을 탐구하였다. 자폐아의 상동적 행동을 식별하기 위해 CNN-GRU 통합 딥 러닝 아키텍처를 활용하였으며, 비통제된 비디오 데이터에서 복잡한 행동 패턴을 정확하게 분류 및 예측하도록 설계되었다. 연구 데이터셋은 YouTube, TikTok 등의 소셜 네트워크 서비스에서 수집된 영상과 Self-Stimulatory Behavior Dataset(SSBD)을 포함하고 있다. 본 모델은 테스트 데이터셋에서 약 62%의 정확도를 달성하였다. 이 결과는 전문 지식이 없는 이들도 자폐 증상을 조기에 인지할 수 있는 가능성을 시사한다. This investigation probed the feasibility of leveraging computer vision and machine learning methodologies for the precocious diagnosis of Autism Spectrum Disorder (ASD). An integrated deep learning framework, amalgamating Convolutional Neural Networks (CNN) with Gated Recurrent Units (GRU), was utilized to facilitate the precise classification and prognostication of intricate behavioral patterns within unstructured video data. The analytical dataset amalgamated footage procured from social networking platforms such as YouTube and TikTok, in conjunction with the Self-Stimulatory Behavior Dataset (SSBD). The model manifested an accuracy rate of approximately 62% on the validation dataset. These findings illuminate the potential for non-experts to discern early indicators of autism, underscoring the transformative implications of this technology in the realm of early ASD intervention.