• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1991년도 추계학술대회 논문집 학회본부
• /
• pp.189-192
• /
• 1991

This paper presents a modular structure design of analog Hopfield neural network. Each multiplier consists of four MOS transistors which are connected to an op-amp at the front end of a neuron. A pair of MOS transistor is used in order to maintain linear operation of the synapse and can produce positive or negative synaptic weight. This architecture can be expandable to any size neural network by forming tree structure. By altering the connections, other nework paradigms can also be implemented using this basic modules. The stength of this approach is the expandability and the general applicability. The layout design of a four-neuron fully connected feedback neural network is presented and is simulated using SPICE. The network shows correct retrival of distorted patterns.

• 정보학연구
• /
• 제4권4호
• /
• pp.51-63
• /
• 2001

잉여수계(Residue Number System)는 각 모듈러스에 자리올림수의 전달이 필요 없고, 병렬 구조를 이루므로, 디지털 신호처리 및 신경망 처리기와 같은 전용 프로세서 설계에서 사용된다. 그러나, 크기 비교 및 부호 검출시에 혼합기수변환(Mixed Radix Conversion)이 요구되며, 이는 전체 연산 속도를 저해하는 요인이 된다. 그러므로 본 논문에서는 혼합기수 변환의 속도를 향상시키기 위하여 잉여수 분할 방법을 개선한 혼합기수변환기를 설계하였다. 설계된 변환기는 기존의 변환기에 비하여 연산기의 크기는 증가하지만 연산시간은 최대 2배가 향상되었다.

• 정보학연구
• /
• 제4권3호
• /
• pp.155-163
• /
• 2001

디지털 신경회로망의 구현에 있어 시그모이드 함수의 구현은 매우 복잡하고 구현하기 어렵다. 따라서, 본 논문에서는 디지털 신경회로망 구현에 문제가 되는 시그모이드 함수처리를 위한 설계 방법을 제안하였다. 제안된 방법은 잉여수계를 이용하여 MAC(Multiplier and Accumulator) 연산 시, 캐리 전파 없이 고속의 연산을 수행할 수 있고 시그모이드 함수처리를 고속으로 수행할 수 있다. 모의실험결과, 각각의 신경 프로세스에 있어서 4.6nsec 이상의 속도를 보임으로써 고속디지털 신경회로망 구현에 적용될 수 있을 것으로 기대된다.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2002년도 ITC-CSCC -2
• /
• pp.790-793
• /
• 2002

In this paper, dynamic path planning of two mobile robots using a modified Hopfield neural network is studied. An area which excludes obstacles and allows gradually changing of activation level of neurons is derived in each step. Next moving step can be determined by searching the next highest activated neuron. By learning repeatedly, the steps will be generated from starting to goal points. A path will be constructed from these steps. Simulation showed the constructed paths of two mobile robots, which are moving across each other to their goals.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2005년도 학술대회 논문집 정보 및 제어부문
• /
• pp.38-40
• /
• 2005

Microdrive with high precision and light mass enough to install on mouse head was fabricated for recording the reliable signal of neuron cell to understand the brain study. The proposed microdrive has three H-form PZT actuators and its guide structure. The microdrive operation principle is based on the well known inchworm principle. The synchronization of three PZT actuators is able to produce the linear motion along the guide structure. Our proposed microdrive has a precise accuracy of about 100nm and a long stroke of about 5mm. The electrode which is used for the recording of the action potential of the neuron cell was fixed at one of PZT actuators. The proposed microdrive was suited to acquisition of signals from in vivo extra-cellular single-unit recoding. On the condition of the anesthetized mouse, the single-unit signals could be recorded by using the proposed microdrive. In addition, applying the PZT microdrive to an alert mouse, we try to implant it on a mouse brain skull to explore single neuron firing.

• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2016년도 추계학술대회
• /
• pp.101-104
• /
• 2016

본 연구에서는 뉴런이 지니는 기능 및 결합구조를 모사하여 3x3 배열의 기능별로 분리시킨 후 디지털회로에서 동작 중 발생할 수 있는 일시적 또는 영구적인 오류 위치를 정확히 찾아내어 복구 시키는 알고리즘을 제안한다. 결합된 세포에서 어느 특정 일부분이 문제가 발생할 경우 그 기능을 다른 세포로 분화되어 동일 기능을 수행하며 오류가 발생한 세포는 주변 세포에 의해 사멸시키는 단계를 거친다. 이런 세포가 지니는 기능 및 구조를 디지털 회로내부에 기능 블록구조로 설계하여 알고리즘을 제안한다.

• 한국전기전자재료학회논문지
• /
• 제35권4호
• /
• pp.311-321
• /
• 2022

The report reviews recent research efforts in demonstrating a computing system whose operation principle mimics the dynamics of biological neurons. The temporal variation of the membrane potential of neurons is one of the key features that contribute to the information processing in the brain. We first summarize the neuron models that explain the experimentally observed change in the membrane potential. The function of ion channels is briefly introduced to understand such change from the molecular viewpoint. Dedicated circuits that can simulate the neuronal dynamics have been developed to reproduce the charging and discharging dynamics of neurons depending on the input ionic current from presynaptic neurons. Key elements include volatile memristors that can undergo volatile resistance switching depending on the voltage bias. This behavior called the threshold switching has been utilized to reproduce the spikes observed in the biological neurons. Various types of threshold switch have been applied in a different configuration in the hardware demonstration of neurons. Recent studies revealed that the memristor-based circuits could provide energy and space efficient options for the demonstration of neurons using the innate physical properties of materials compared to the options demonstrated with the conventional complementary metal-oxide-semiconductors (CMOS).

• 대한수의학회지
• /
• 제42권4호
• /
• pp.537-543
• /
• 2002

The purpose of the present study was to investigate structural stability of extensive laminectomy and the effect of subcutaneous fat autograft on restricting formation of postlaminectomy membrane (fibrous tissue). Eighteen healthy dogs of both sexes and of mixed breeding were divided into 6 groups : (1) unilateral hemilaminectomy (group H) on 3rd, 4th and 7th vertebrae ; (2) modified dorsal laminectomy (group D) on 3rd, 4th and 7th vertebrae. Hemilaminectomies were carried out incontinuously at 16 sites in 4 dogs, then subcutaneous fat autografts (group F) were applied to 8 laminectomy sites and no treatment (group C) was assigned to 8 laminectomy sites, too. Operating time of group H ($30.9{\pm}10.4$ minutes) was significantly shorter (p<0.05) than that of group D ($43.1{\pm}12.2$ minutes), but surgical hemorrhage of group H is severer than that of group D. General states, such as standing, gait, defecation and urination, were normal. Upper motor neuron/lower motor neuron signs were not found and superficial/deep pain, proprioception and anal sphincter tone were normal. Gross postmortem findings were similar in all groups. The laminectomy sites of groups H and D were filled with fibous connective tissue at 4 months after operation and histopathological abnormalities of spinal cord were not found. One of eight laminectomy sites in group F was filled with fibrous tissue at 2 months after operation, but all operating sites of group C were covered with fibrous tissue. The present study indicated that extensive laminectomy on 7 vertebrae, using unilateral hemilaminectomy and modified dorsal laminectomy technique, maintained structural stability and subcutaneous fat autograft was effective on reducing the formation of fibrous membrane in laminectomy sites.

• 조명전기설비학회논문지
• /
• 제25권9호
• /
• pp.50-55
• /
• 2011

This research is to apply the adaptive control of neuron networks for the real-time attitude control of Multi-articulated robot. Multi-articulated robot is expressed with a complicated mathematical model on account of the mechanic, electric non-linearity which each articulation of mechanism has, and includes an unstable factor in time of attitude control. If such a complex expression is included in control operation, it leads to the disadvantage that operation time is lengthened. Thus, if the rapid change of the load or the disturbance is given, it is difficult to fulfill the control of desired performance. In this research we used the response property curve of the robot instead of the activation function of neural network algorithms, so the adaptive control system of neural networks constructed without the information of modeling can perform a real-time control. The proposed adaptive control algorithm generated control signs corresponding to the non-linearity of Multi-articulated robot, which could generate desired motion in real time.

• 조명전기설비학회논문지
• /
• 제26권10호
• /
• pp.89-94
• /
• 2012

This research is to apply the control of neuron networks for the real-time walking control of Multi-articulated robot. Multi-articulated robot is expressed with a complicated mathematical model on account of the mechanic, electric non-linearity which each articulation of mechanism has, and includes an unstable factor in time of walking control. If such a complex expression is included in control operation, it leads to the disadvantage that operation time is lengthened. Thus, if the rapid change of the load or the disturbance is given, it is difficult to fulfill the control of desired performance. This paper proposes a new mode to implement a neural network controller by installing a real object for controlling and an algorithm for this, which can replace the existing method of implementing a neural network controller by utilizing activation function at the output node. The proposed control algorithm generated control signs corresponding to the non-linearity of Multi-articulated robot, which could generate desired motion in real time.