http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Implementation of communication system using signals originating from facial muscle constructions
EungSoo Kim,TaeWan Eum 한국지능시스템학회 2004 INTERNATIONAL JOURNAL of FUZZY LOGIC and INTELLIGE Vol.4 No.2
A person does communication between each other using language. But, In the case of disabled person, cannot communicate own idea to use writing and gesture. We embodied communication system using the EEG so that disabled person can do communication. After feature extraction of the EEG included facial muscle signals, it is converted the facial muscle into control signal, and then did so that can select character and communicate idea.
Dynamics of neural net model with arbitry connectivity and time delay
Kim, EungSoo 선문대학교 첨단과학기술연구소 1996 첨단과학기술연구소 논문집 Vol.1 No.1
지연층으로부터 임의의 연결도를 가지는 계열연상 신경회로망 모델의 다이나믹스에 관하여 논한다. 다양한 연결도의 변화로부터 신경회로망의 특성이 변해감을 수치실험을 통해 확인할 수 있었다. 부하율로 볼 때 임의의 연결도 c가 감소함에 따라 메모리 용량은 증가하는데, 이것은 부하율의 관점에서 볼 때 회로망이 낮은 연결도에서 잘 동작한다는 것을 나타낸다. 한편, 상기특성은 c가 감소함에 따라 감소하는 특성을 보인다. 이렇듯 다양한 연결도를 가지는 지연 신경회로망의 계열상기특성을 통계학적 신경역학 방법으로 해석하였다. 이론에서 도출된 결과는 수치실험에서 나타난 여러 가지 현상들을 잘 설명해주고 있다. The dynamics of a sequential associative neural network model with arbitrary connectivity from delay layer arc discussed. For the varied arbitrary connectivity, I can find various characteristics from the simulations. The memory capacity increases as arbitrary connectivity c is decreased in the sense of the loading ratio. This say that the network works better with lower connectivity in the sense of the loading ratio (the loading ratio represents a loading per connection). The retrieval ability of the network decrease as c is decreased. The method of statistical neurodynamical method is used for analyzing the retrieval dynamics of a sequential associative neural network model with multi delayed synapses which has a various connectivity. The theory explains dynamical behaviors in retrieval processes which arc observed by computer simulations. The method is based only on probability and approximation calculations also. The theory explains the non-trivial behavior of the network observed in computer simulations.
신응수(Eungsoo Shin),이종화(Jongwha Lee) 대한기계학회 2003 대한기계학회 춘추학술대회 Vol.2003 No.4
This study intends to provide the analytical and experimental damping characterization of carbon<br/> nanotube/epoxy composites. A constitutive model based on continuum mechanics is employed to describe<br/> epoxy and the perfectly bonded and partially bonded nanotubes. An interfacial stick-slip between the<br/> nanotubes and epoxy is considered to characterize the damping of the composites. For experimental<br/> estimation, beam-type specimens are prepared with a variation of nanotube concentration from 0.5% to 2% in<br/> weight. An ultrasonic agitation method is employed for enhancing the nanotube dispersion within epoxy.<br/> Damping of the composites is characterized in terms of the strain and the nanotube concentration. Results<br/> show that the nanotube concentration significantly affects the damping characteristics of the nanocomposites.<br/> A good correlation is found between the analytical prediction based on the stick-slip and the experimental<br/> measurements.
Implementation of communication system using signals originating from facial muscle constructions
Kim, EungSoo,Eum, TaeWan Korean Institute of Intelligent Systems 2004 INTERNATIONAL JOURNAL of FUZZY LOGIC and INTELLIGE Vol.4 No.2
A person does communication between each other using language. But, In the case of disabled person, cannot communicate own idea to use writing and gesture. We embodied communication system using the EEG so that disabled person can do communication. After feature extraction of the EEG included facial muscle signals, it is converted the facial muscle into control signal, and then did so that can select character and communicate idea.
Relationship between various weight strengths and association capabilities of the neural network
Kim, EungSoo 선문대학교 ·중소기업기술지원연구소 1996 선문공대 연구/기술 논문집 Vol.1 No.1
신경회로망의 이론적 해석에 있어서 지금까지는 주로 고정된 파라메터값들을 지닌 경우에 대해 그 특성들이 조사 연구되었다. 그러나 신경회로망의 가장 중요한 특징 중의 하나가 스스로 환경에 적응해 가는 능력에 있음을 인식한다면, 균일하지 않은 여러 가지 상황으로 신경망을 구성하여 그 때의 동작이나 능력을 평가하고 이해하는 것은 신경망 연구에 있어서 대단히 중요한 접근이다. 이러한 입장에서 본 논문에서는 지연층을 가지며 가변적 시냅스 값을 가진 신경회로망을 대상으로 그 능력을 해석적으로 평가한다. Theoretical studies on neural network models focus very often on the situation where relevant parameters are fixed precisely. For example, the thresholds and/or the output amplitudes of neurons, or the synaptic connections, are precise. To discuss the robustness of neural networks, which is the most important characteristic of adaptive systems, I have to take various non-uniformities into account, and see what happens to theoretical results in an ideal situation. The nature of neural network systems could only be understood by studying various kind of trade-offs. With these things in mind, I study the dynamics of an sequential associative neural network model with varied weight strength from delay layer.
E-clutch system에서 변속 의지 판단 알고리즘에 대한 연구
이응수(Eungsoo Lee),이승호(Seungho Lee),성윤현(Yunhyun Sung) 한국자동차공학회 2018 한국자동차공학회 학술대회 및 전시회 Vol.2018 No.11
The 2 pedal E-clutch system is based on a manual clutch system that eliminates the clutch pedal and automatically engages and disengages the clutch upon gear change. The advantage of this system is that even the driver who is inexperienced can easily shift gears only by operating the gear lever and automatically releases the clutch when the coasting conditions are satisfied, thereby contributing to fuel efficiency improvement. In addition, without worrying about turning off the engine, the system can smoothly start the vehicle without hitting the clutch pedal, even during hill climbing or traffic congestion. In this study, we developed a software to shift intend algorithm based on the driver"s gear lever motion in a 2-pedal E-clutch system and verified in Demo-car.