• 한국산업정보학회논문지
• /
• 제4권1호
• /
• pp.68-76
• /
• 1999

신경회로망을 동적 정보처리에 응용하기 위해서는 비대칭 결합 신경회로망에서 생성되는 동적 상태천이에 관한 직관적 이해가 필요하다. 본 논문에서는 각 뉴런이 최근접 뉴런에만 양자화 결합하중 +1및 -1로 연결된 환상형 신경회로망의 동적인 상태천이 특성을 해석하였다. 상태천이 해석 알고리즘을 이용한 시뮬레이션 결과, 네트워크는 고정점, 베이슨을 갖는 리미트사이클 및 베이슨이 없는 리미트사이클의 3가지 어트랙터를 가진다. 또한, 네트워크에서 생성되는 리미트사이클의 수와 주기를 이론적으로 해석하여 정식화하고, 리미트사이클을 구성하는 상태벡터의 필요조건을 나타내었다. 이론 해석의 결과는 네트워크에서 생성되는 리미트사이클의 수가 뉴런(소자)의 수 n에 대해서 지수 함수적으로 증가함을 보여준다. 따라서 순환결합형 신경회로망은 많은 동적 정보를 리미트사이클로 저장하는 메모리 시스템으로 이용할 수 있다.

• 한국산학기술학회논문지
• /
• 제9권6호
• /
• pp.1574-1580
• /
• 2008

본 논문에서는 외란이나 시스템의 파라미터 변동 및 불확실성 등이 존재하는 자동화 설비시스템을 강인하고 정밀하게 제어할 수 있도록 하기 위해 동적 신경망 처리기(DNP)인 신경망 제어기를 설계하였다. 자동화 설비시스템에서 부품의 조립, 가공 등 복잡하고 정교한 임무를 수행시키기 위해서는 end-effector의 이동경로 궤적에 대한 추적제어 뿐만 아니라 목표물에 대하여 접촉하는 힘의 궤적에 대한 추적 제어가 필수적이다. 또한 자동화 설비시스템에서 플랜트의 역기구학적인 좌표변환을 계산하기 위한 학습구조를 개발하였으며, DNP가 이용될 수 있는 예를 설명하였다. 제안된 동적 신경망인 DNP의 구조와 학습 알고리즘을 제시하고 컴퓨터 모의실험을 통해 학습 성능을 증명하였다.

• 대한전기학회논문지:시스템및제어부문D
• /
• 제53권1호
• /
• pp.22-31
• /
• 2004

This paper deals with the dynamic hand gesture recognition based on computer vision using neural networks. This paper proposes a global search method and a local search method to recognize the hand gesture. The global search recognizes a hand among the hand candidates through the entire image search, and the local search recognizes and tracks only the hand through the block search. Dynamic hand gesture recognition method is based on the skin-color and shape analysis with the invariant moment and direction information. Starting point and ending point of the dynamic hand gesture are obtained from hand shape. Experiments have been conducted for hand extraction, hand recognition and dynamic hand gesture recognition. Experimental results show the validity of the proposed method.

• Journal of Power Electronics
• /
• 제5권1호
• /
• pp.36-44
• /
• 2005

This paper presents an implementation of high performance control of a reluctance synchronous motor (RSM) using a neural network with a direct torque control. The equivalent circuit in a RSM, which considers iron losses, is theoretically analyzed. Also, the optimal current ratio between torque current and exiting current is analytically derived. In the case of a RSM, unlike an induction motor, torque dynamics can only be maintained by controlling the flux level because torque is directly proportional to the stator current. The neural network is used to efficiently drive the RSM. The TMS320C3l is employed as a control driver to implement complex control algorithms. The experimental results are presented to validate the applicability of the proposed method. The developed control system shows high efficiency and good dynamic response features for a 1.0 [kW] RSM having a 2.57 ratio of d/q.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1991년도 하계학술대회 논문집
• /
• pp.333-336
• /
• 1991

In this paper, Hopfield & Tank model-like artificial neural network structure is proposed, which can be used for the optimal path planning problems such as the unit commitment problems or the maintenance scheduling problems which have been solved by the dynamic programming method or the branch and bound method. To construct the structure of the neural network, an energy function is defined, of which the global minimum means the optimal path of the problem. To avoid falling into one of the local minima during the optimization process, the simulated annealing method is applied via making the slope of the sigmoid transfer functions steeper gradually while the process progresses. As a result, computer(IBM 386-AT 34MHz) simulations can finish the optimal unit commitment problem with 10 power units and 24 hour periods (1 hour factor) in 5 minites. Furthermore, if the full parallel neural network hardware is contructed, the optimization time will be reduced remarkably.

• 한국정밀공학회지
• /
• 제25권2호
• /
• pp.140-147
• /
• 2008

This paper describes a Neural Network based PD control scheme for motion control of pneumatic servo cylinder. Pneumatic systems have inherent nonlinearities such as compressibility of air and nonlinear frictions present in cylinder. The conventional linear controller is limited in some applications where the affection of nonlinear factor is dominant. A self-excited oscillation method is applied to derive the dynamic design parameters of linear model. Based on the parameters thus identified, a PD feedback compensator is designed first and then a neural network is incorporated. The experiments of a trajectory tracking control using the proposed control scheme are performed and a significant reduction in tracking error is achieved by comparing with those of a PD control.

• Transactions on Control, Automation and Systems Engineering
• /
• 제3권4호
• /
• pp.272-282
• /
• 2001

Uncertainties are the main reasons of deterioration of contour control of industrial articulated robot arm. In this paper, a high-precision contour control method was proposed to overcome some main uncertainties, such as torque saturation, system delay dynamics, interference between robot links, friction, and so on. Firstly, each considered factor of uncertainties was introduced briefly. Then proper realizable objective trajectory generation was presented to avoid torque saturation from objective trajectory. According to the model of industrial articulated robot arm, construction of Gaussian neural network controller with considering system delay dynamic, interference between robot links and friction was explained in detail. Finally, through the experiment and simulation, the effectiveness of proposed method was verified. Furthermore, based on the results it was shown that the Gaussian neural network controller can be also adapted for the various kinds of friction and high-speed motion of industrial articulated robot arm.

• 한국정밀공학회지
• /
• 제14권7호
• /
• pp.29-38
• /
• 1997

A closed-loop system modeling of an active/semiactive suspension system has been accomplished through an artificial neural network. A 7DOF full model as a system's equation of motion has been derived and an output feedback linear quadratic regulator has been designed for control purpose. A training set of a sample data has been obtained through a computer simulation. A 7DOF full model with LQR controller simulated under several road conditions such as sinusoidal bumps and rectangular bumps. A general multilayer perceptron neural network is used for dynamic modeling and target outputs are fedback to the a layer. A backpropagation method is used as a training algorithm. Model validation of new dataset have been shown through computer simulations.

• 제어로봇시스템학회논문지
• /
• 제4권5호
• /
• pp.651-657
• /
• 1998

In this study, an intelligent scheduling with hybrid architecture, which integrates expert system and neural network, is proposed. Neural network is trained with the data acquired from simulation model of FMC to obtain the knowledge about the relationship between the state of the FMC and its best dispatching rule. Expert system controls the scheduling of FMC by integrating the output of neural network, the states of FMS, and user input. By applying the hybrid system to a scheduling problem, the human knowledge on scheduling and the generation of non-logical knowledge by machine teaming, can be processed in one scheduler. The computer simulation shows that comparing with MST(Minimum Slack Time), there is a little increment in tardness, 5% growth in flow time. And at breakdown, tardness is not increased by expert system comparing with EDD(Earliest Due Date).

• 한국정보과학회:학술대회논문집
• /
• 한국정보과학회 2006년도 가을 학술발표논문집 Vol.33 No.2 (B)
• /
• pp.218-222
• /
• 2006

우수한 능력의 인공지능 개체로 구성된 게임은 그렇지 못한 게임에 비해 더 나은 흥미를 사용자에게 제공할 수 있다. 미국 Valve사의 Half-Life, Counter-Strike 및 한국 Dragonfly사의 Special-Force와 같은 실시간 FPS 전투게임에서 상대편에 대한 검색 및 목표 화하는(Targeting) 기법은 인공개체의 전투력에 중요한 하나의 요소이다. 하지만 이 같은 경우의Targeting은 정적인 대상에 대한 것이 아니라 동적인 대상에 대한 것이므로 단순한 산술 계산으로는 실용적인 효과를 내기 힘들다. 본 논문에서는 Neural Network를 이용한 학습기법을 사용하여 동적인 개체에 대한 효과적인 Targeting기법을 제안한다. 제안한 기법은 매 순간 변화하는 상황정보와 Virtual bullet이라는 가상 미사일 개념을 활용하여 학습 Data를 모델링한 후 Neural Network로 학습시켜 효과적인 Targeting이 가능하도록 구현하였다. 제안한 기법은 Java기반의 탱크전투 시뮬레이션 Framework인 Robocode에 적용하여 그 성능을 평가하였다. 제안된 기법으로 제작된 Robot(Crystal 1.0)은 ‘2006 Robocode Korea Cup에서 우승을 차지하였다.