• 전기학회논문지
• /
• 제66권12호
• /
• pp.1772-1781
• /
• 2017

In this study, the design methodology for alleviating the overfitting problem of Polynomial Neural Networks(PNN) is realized with the aid of two kinds techniques such as L2 regularization and Sum of Squared Coefficients (SSC). The PNN is widely used as a kind of mathematical modeling methods such as the identification of linear system by input/output data and the regression analysis modeling method for prediction problem. PNN is an algorithm that obtains preferred network structure by generating consecutive layers as well as nodes by using a multivariate polynomial subexpression. It has much fewer nodes and more flexible adaptability than existing neural network algorithms. However, such algorithms lead to overfitting problems due to noise sensitivity as well as excessive trainning while generation of successive network layers. To alleviate such overfitting problem and also effectively design its ensuing deep network structure, two techniques are introduced. That is we use the two techniques of both SSC(Sum of Squared Coefficients) and $L_2$ regularization for consecutive generation of each layer's nodes as well as each layer in order to construct the deep PNN structure. The technique of $L_2$ regularization is used for the minimum coefficient estimation by adding penalty term to cost function. $L_2$ regularization is a kind of representative methods of reducing the influence of noise by flattening the solution space and also lessening coefficient size. The technique for the SSC is implemented for the minimization of Sum of Squared Coefficients of polynomial instead of using the square of errors. In the sequel, the overfitting problem of the deep PNN structure is stabilized by the proposed method. This study leads to the possibility of deep network structure design as well as big data processing and also the superiority of the network performance through experiments is shown.

• 한국통신학회논문지
• /
• 제28권10C호
• /
• pp.1023-1032
• /
• 2003

MPEG-4와 같은 객체 기반 부호화는 멀티미디어 응용을 위한 다양한 내용 기반 기능들을 제공한다. 압축 효율의 향상과 더불어 이러한 기능들이 지원되도록 하기 위해서는 비디오 데이터의 각 프레임은 비디오 객체로 분할되어야 한다. 본 논문에서는 비선형 다중스케일 필터링과 시공간 정보를 사용한 효과적인 비디오 객체 분할 기법을 제안한다. 제안된 방법은 안정화된 역 확산 방정식(Stabilized Inverse Diffusion Equation : SIDE)에 기반한 비선형 다중스케일 필터링을 사용하여 공간적 분할을 수행한다. 또한 구해진 초기 분할된 영역들은 인접 영역 그래프 (Region Adjacency Graph : RAG)를 사용하여 병합된다. 본 논문에서는 통계적 유의성 검사(Statistical significance test)와 시변 메모리(Time-variant memory)를 시간적 분할 방법으로 사용하며 구해진 공간적 분할과 시간적 분할을 결합하여 최종 객체 영역을 효과적으로 분할한다. 본 논문에서 제안된 공간적 분할 방법은 기존의 형태학적 Watershed 알고리즘에 비해 잡음에 강인한 분할 특성을 나타내었으며 기존의 A. Neri의 방법과 비교하였을 때, 최종 분할된 객체 영역의 정확도 비율이 Akiyo는 43%, Claire는 29% 정도 향상됨을 확인할 수 있었다.

• International Journal of Control, Automation, and Systems
• /
• 제3권4호
• /
• pp.524-532
• /
• 2005

This paper concerns delay-dependent guaranteed cost control (GCC) problem for a class of linear state-delayed systems with norm-bounded time-varying parametric uncertainties. By incorporating the free weighing matrix approach developed recently, new delay-dependent conditions for the existence of the guaranteed cost controller are presented in terms of matrix inequalities for both nominal state-delayed systems and uncertain state-delayed systems. An algorithm involving convex optimization is proposed to design a controller achieving a suboptimal guaranteed cost such that the system can be stabilized for all admissible uncertainties. Through numerical examples, it is shown that the proposed method can yield less guaranteed cost than the existing delay-dependent methods.

• 한국전산유체공학회지
• /
• 제12권1호
• /
• pp.35-42
• /
• 2007

This paper describes a recent development on the divergence free basis function based on a hermite stream function and verifies its validity by comparing results with those from a modified residual method known as one of stabilized finite element methods. It can be shown that a proper choice of degrees of freedom at a node with a proper arrangement of the hermite interpolation functions can yield solenoidal or divergent free interpolation functions for the velocities. The well-known cavity problem has been chosen for validity of the present algorithm. The comparisons from numerical results between the present and the modified residual showed the present method yields better results in both the velocity and the pressure within modest Reynolds numbers(Re = 1,000).

• 한국전산유체공학회:학술대회논문집
• /
• 한국전산유체공학회 2009년 추계학술대회논문집
• /
• pp.49-55
• /
• 2009

This paper evaluates performances of a recently developed divergence-free finite element method based on Hermite interpolated stream functions. Velocity bases are derived from Hermite interpolated stream functions to form divergence-free basis functions. These velocity basis functions constitute a solenoidal function space, and the simple gradient of the Hermite functions constitute an irrotational function space. The incompressible Navier-Stokes equation is orthogonally decomposed into a solenoidal and an irrotational parts, and the decoupled Navier-Stokes equations are projected onto their corresponding spaces to form proper variational formulations. To access accuracy and convergence of the present algorithm, three test problems are selected. They are lid-driven cavity flow, flow over a backward-facing step and buoyancy-driven flow within a square enclosure. Hermite interpolation functions from cubic to quintic are chosen to run the test problems. Numerical results are shown. In all cases it has shown that the present method has performed well in accuracies and convergences. Moreover, the present method does not require an upwinding or a stabilized term.

• 한국군사과학기술학회지
• /
• 제16권3호
• /
• pp.233-239
• /
• 2013

This paper includes the algorithm of eliminating noise, the processing technique of sensor and the results of designing vital and environmental sensor, one of the survivability subsystem of Future Soldier System. In this paper, we propose the adaptive filtering, moving noise removal in order to detect signals stabilized. And these help that we get bio-signals the ECG calculating methods such as search back and ensemble method. It is made up the vital and environmental sensor including the flexible sensor. In that sense, this study can be applied when it is planned the modular type Future Soldier System.

• 한국군사과학기술학회지
• /
• 제2권1호
• /
• pp.150-158
• /
• 1999

본 논문은 선박 장착을 위한 안정된 플랫폼형의 동적방식 광파이버 자이로콤파스의 구동시스템 안정화에 대한 연구이다. 구동모터로 선정된 스텝핑모터는 FOG 콤파스 특성에 따른 연구 목적상 정속 및 저진동의 안정한 운전을 필요로 한다. 이러한 측면에서 기발표된 주파수변조방식의 안정화 구동시스템의 설계상 나타났던 문제점을 보완할 수 있는 퍼지제어를 사용한 새로운 방식의 제어기를 제안한다. 시뮬레이션을 통하여 제안된 방식의 우수성을 보였으며, 특히 시스템의 크기, 무게 그리고 운전환경의 다양성 측면에서 자이로콤파스로서의 적용 가능성이 크다.

• Transactions on Control, Automation and Systems Engineering
• /
• 제1권2호
• /
• pp.128-133
• /
• 1999

This paper presents a new delay-dependent robust stabilization condition for uncertain time-delay systems. An algorithm involving convex optimization is proposed to compute a suboptimal upper bound of the delay such that the system can be stabilized by the controller for all admissible uncertainties. It is illustrated by numerical examples that the proposed delay-dependent controller can be less conservative than previous results. It is also shown that the proposed delay-dependent controller can even capture the delay-independent stability of the system, which is not possible with existing delay-dependent results.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2005년도 ICCAS
• /
• pp.241-245
• /
• 2005

Strapdown inertial navigation system (SINS) integrated with astronavigation system (ANS) yields reliable mission capability and enhanced navigational accuracy for spacecrafts. The theory and characteristics of integrated system based on unscented Kalman filtering is investigated in this paper. This Kalman filter structure uses unscented transform to approximate the result of applying a specified nonlinear transformation to a given mean and covariance estimate. The filter implementation subsumed here is in a direct feedback mode. Axes misalignment angles of the SINS are observation to the filter. A simple approach for simulation of axes misalignment using stars observation is presented. The SINS error model required for the filtering algorithm is derived in space-stabilized mechanization. Simulation results of the integrated navigation system using a medium accuracy SINS demonstrates the validity of this method on improving the navigation system accuracy with the estimation and compensation for gyros drift, and the position and velocity errors that occur due to the axes misalignments.

• 전력전자학회:학술대회논문집
• /
• 전력전자학회 1999년도 전력전자학술대회 논문집
• /
• pp.93-96
• /
• 1999

Solar cell generate DC power from sunlight whose power is different at any instance according to condition of variables : insolation and temperature. In order to improve the system utility factor and efficiency of energy conversion, it is desirable to operate the PV system at maximum power point of solar cell under different condition. In this paper, Boost chopper is controlled it output voltage with a new discrete control algorithm for MPPT. PWM signal of DC-DC converter are generated with a 89C51 microcontroller. Switching frequency of DC-DC converter is set at 10KHz. Simulation and experimental results show that the PV system studied in this paper is always operated at maximum power point under different maximum power point of solar cells having stabilized output voltage waveform with relatively small ripple component