• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2009년도 정보 및 제어 심포지움 논문집
• /
• pp.405-407
• /
• 2009

SVM(Support Vector Machine) is a classification method which is recently watched in mechanical learning system. Vapnik, Osuna, Platt etc. had suggested methodology in order to solve needed QP(Quadratic Programming) to realize SVM so that have extended application field. SVM find hyperplane which classify into 2 class by converting from input space converter vector to characteristic space vector using Kernel Function. This is very systematic and theoretical more than neural network which is experiential study method. Although SVM has superior generalization characteristic, it depends on Kernel Function. There are three category in the Kernel Function as Polynomial Kernel, RBF(Radial Basis Function) Kernel, Sigmoid Kernel. This paper has analyzed performance of SVM against kernel using virtual data.

• 한국해양공학회지
• /
• 제23권5호
• /
• pp.1-8
• /
• 2009

This paper is a continuation of the recent development on Hermite-based divergence free element method and deals with a non-isothermal fluid flow thru the buoyancy driven flow in a square enclosure with temperature difference across the two sides. The basis functions for the velocity field consist of the Hermite function and its curl while the basis functions for the temperature field consists of the Hermite function and its gradients. Hence, the number of degrees of freedom at a node becomes 6, which are the stream function, two velocities, the temperature and its x and y derivatives. This paper presents numerical results for Ra = 105, and compares with those from a stabilized finite element method developed by Illinca et al. (2000). The comparison has been done on 32 by 32 uniform elements and the degree of approximation of elements used for the stabilized finite element are linear (Deg. 1) and quadratic (Deg. 2). The numerical results from both methods show well agreements with those of De vahl Davi (1983).

• 한국소음진동공학회논문집
• /
• 제12권1호
• /
• pp.12-20
• /
• 2002

To control the noise of an enclosed sound field, we built a state space model using the acoustic modal parameter description. Using the state space model, we can investigate the controllability and observability, and find an appropriate position of control speaker and microphone to control sound field of the enclosed space. We implemented LQR(linear quadratic regulator) controller and reduced order observer to reduce the first acoustic mode. Experiments showed satisfactory results of 4∼10 dB reduction of magnitude of the first acoustic mode, and support the feasibility of the proposed scheme to lightly damped acoustic field.

• 대한조선학회논문집
• /
• 제40권5호
• /
• pp.26-35
• /
• 2003

Propellers operating in a given nonuniform ship wake generate unsteady loads leading to undesirable stern vibration problems. The skew is known to be the most proper and effective geometric parameter to control or reduce the fluctuating forces on the shaft. This paper assumes the skew profile as either a quadratic or a cubic function of the radius and determines the coefficients of the polynomial function by applying the simplex method. The method uses the converted unconstrained algorithm to solve the constrained minimization problem of 6-component shaft excitation forces. The propeller excitation was computed either by applying the two-dimensional gust theory for quick estimation or by the fully three-dimensional unsteady lifting surface theory in time domain for an accurate solution. A sample result demonstrates that the shaft forces can be further reduced through optimization from the original design.

• 대한수학회보
• /
• 제54권6호
• /
• pp.2165-2182
• /
• 2017

Let G be a group and $\mathbb{C}$ the field of complex numbers. Suppose ${\sigma}:G{\rightarrow}G$ is an endomorphism satisfying ${{\sigma}}({{\sigma}}(x))=x$ for all x in G. In this paper, we first determine the central solution, f : G or $G{\times}G{\rightarrow}\mathbb{C}$, of the functional equation $f(xy)+f({\sigma}(y)x)=2f(x)+2f(y)$ for all $x,y{\in}G$, which is a variant of the quadratic functional equation. Using the central solution of this functional equation, we determine the general solution of the functional equation f(pr, qs) + f(sp, rq) = 2f(p, q) + 2f(r, s) for all $p,q,r,s{\in}G$, which is a variant of the equation f(pr, qs) + f(ps, qr) = 2f(p, q) + 2f(r, s) studied by Chung, Kannappan, Ng and Sahoo in [3] (see also [16]). Finally, we determine the solutions of this equation on the free groups generated by one element, the cyclic groups of order m, the symmetric groups of order m, and the dihedral groups of order 2m for $m{\geq}2$.

• 한국정보통신학회논문지
• /
• 제4권3호
• /
• pp.697-703
• /
• 2000

임팩트이온화현상은 소자의 크기가 점점 작아지면서, 높은 에너지에 있는 hot carrier 전송 을 해석하기 위해 매우 중요하므로 소자의 시뮬레이션에 정확한 임팩트이온화모델이 필수적이다. 털 연구에서는 의사포텐셜방법을 사용하여 풀밴드모델을 구하고, 임팩트이온화율은 수정된 Keldysh 공식을 이용하여 유도하였다. 본 연구에서는 Gahs 임팩트이은화의 온도의존특성을 조사하기 위하여 Monte Carlo 시뮬레이터를 제작하여 임팩트이온화계수를 구하였다. 결과적으로, 임팩트이온화계수는 300K에서 실험값과 잘 일치하였다. 또한 에너지는 전계가 증가할수록 증가하고, 높은 온도에서는 포논 산란의 emission mode가 높기 때문에 에너지가 감소함을 알 수 있었다. 임팩트이온화의 대수 fitting 함수 식은 온도와 전계에 대해 2차식으로 표현하였다. 대수 fitting 함수의 오차는 대부분 5%이내에 머물렀다. 그러므로 대수식으로 표현된 임팩트 이온화계수는 온도에 의존함을 알았고, 임팩트이온화계수를 구하는데 시간을 절약할 수 있다.

• 한국정보통신학회:학술대회논문집
• /
• 한국해양정보통신학회 2000년도 춘계종합학술대회
• /
• pp.460-464
• /
• 2000

임팩트이온화현상은 소자의 크기가 점점 작아지면서, 높은 에너지에 있는 hot carrier 전송을 해석하기 위해 매우 중요하므로 소자의 시뮬레이션에 정확한 임팩트이온화모델이 필수적이다. 본 연구에서는 의사포텐셜방법을 사용하여 full 밴드모델을 구하고, 임팩트이온화율은 수정된 Keldysh 공식을 이용하여 유도하였다. 본 연구에서는 GaAs 임팩트이온화의 온도와 전계에 대한 의존특성을 조사하기 위하여 Monte Carlo 시뮬레이터를 제작하여 임팩트이온화계수를 구하였다. 결과적으로, 임팩트이온화계수는 300K에서 실험 값과 잘 일치하였다. 또한 에너지는 전계가 증가할수륵 증가하고, 높은 온도에서는 포논산란의 emission mode가 높기 때문에 에너지가 감소함을 알 수 있었다. 임팩트 이온화의 대수 fitting 함수 식은 온도와 전계에 대해 2차 식으로 표현하였다. 대수 fitting 함수의 오차는 대부분 5%이내에 머물렀다. 그러므로 대수 식으로 표현된 임팩트이온화계수는 온도와 전계에 의존함을 알았고, 온도와 전계에 의존하는 임팩트이온화계수를 구하는데 시간을 절약할 수 있다

• Structural Engineering and Mechanics
• /
• 제4권6호
• /
• pp.703-715
• /
• 1996

In stochastic analysis, the randomness of the structural parameters is taken into consideration and the response variability is obtained in addition to the conventional (mean) response. In the present paper the structural response variability of plate structure is calculated using the weighted integral method and is compared with the results obtained by different methods. The stochastic field is assumed to be normally distributed and to have the homogeneity. The decomposition of strain-displacement matrix enabled us to extend the formulation to the stochastic analysis with the quadratic elements in the weighted integral method. A new auto-correlation function is derived considering the uncertainty of plate thickness. The results obtained in the numerical examples by two different methods, i.e., weighted integral method and Monte Carlo simulation, are in a close agreement. In the case of the variable plate thickness, the obtained results are in good agreement with those of Lawrence and Monte Carlo simulation.

• 한국군사과학기술학회지
• /
• 제12권4호
• /
• pp.540-547
• /
• 2009

In the near future, military robots are likely to be substituted for military personnel in the field of battle. The power system of a legged robot is considerably more complex than the one used for a land vehicle because of the coordination and stability issues due to the large number of degree of freedom. In this paper, a servovalve-piston combination system for a straight-line motion of robot leg is modeled as three degree of freedom based on double inputs and single output transfer function. The output is the displacement of piston from neutral. The inputs are valve displacement from neutral and arbitrary load force in this system. LQR(Linear Quadratic Regulator) technique is applied in order to achieve robust stability and fast responses of the system. The Kalman filter loop, rejection of disturbance and noise, riccati equation, filter gain matrix, and frequency domain equality are analyzed and designed.

• Journal of Radiation Protection and Research
• /
• 제28권1호
• /
• pp.59-63
• /
• 2003

Purpose : Measurement of transmission dose is useful for in vivo dosimetry of QA purpose. The objective of this study is to develope an algorithm for estimation of tumor dose using measured transmission dose for open radiation field. Materials and Methods : Transmission dose was measured with various field size (FS), phantom thickness (Tp), and phantom chamber distance (PCD) with a acrylic phantom for 6 MV and 10 MV X-ray Source to chamber distance (SCD) was set to 150 cm. Measurement was conducted with a 0.6 cc Farmer type ion chamber. Using measured data and regression analysis, an algorithm was developed for estimation of expected reading of transmission dose. Accuracy of the algorithm was tested with flat solid phantom with various settings. Results : The algorithm consisted of quadratic function of log(A/P) (where A/P is area-perimeter ratio) and tertiary function of PCD. The algorithm could estimate dose with very high accuracy for open square field, with errors within ${\pm}0.5%$. For elongated radiation field, the errors were limited to ${\pm}1.0%$. Conclusion : The developed algorithm can accurately estimate the transmission dose in open radiation fields with various treatment settings.