• ETRI Journal
• /
• 제28권1호
• /
• pp.59-66
• /
• 2006

Learning behaviors of a radial basis function network (RBFN) using a singular value decomposition (SVD) and stochastic gradient (SG) algorithm, together named RBF-SVD-SG, for odor sensing systems are analyzed, and a fast training method is proposed. RBF input data is from a conducting polymer sensor array. It is revealed in this paper that the SG algorithm for the fine-tuning of centers and widths still shows ill-behaving learning results when a sufficiently small convergence coefficient is not used. Since the tuning of centers in RBFN plays a dominant role in the performance of RBFN odor sensing systems, our analysis is focused on the center-gradient variance of the RBFN-SVD-SG algorithm. We found analytically that the steadystate weight fluctuation and large values of a convergence coefficient can lead to an increase in variance of the center-gradient estimate. Based on this analysis, we propose to use the least mean square algorithm instead of SVD in adjusting the weight for stable steady-state weight behavior. Experimental results of the proposed algorithm have shown faster learning speed and better classification performance.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2000년도 제15차 학술회의논문집
• /
• pp.200-200
• /
• 2000

Highly nonlinear dynamical systems are easily identified using neural networks. When disturbances are included in the learning data set Int system modeling, modeling process will be poorly performed. Since the radial basis functions in the radial basis function network(RBFN) are centered at the points specified by the weights, RBF networks are robust for approximating the process including the narrow-band disturbances deviating significantly from the regular signals. To exclude(filter) these disturbances, a robust algorithm for system identification, based on the RBFN, is proposed. The performance of system identification excluding disturbances is investigated and compared with the one including disturbances.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2007년도 제38회 하계학술대회
• /
• pp.1938-1939
• /
• 2007

Radial Basis Function Network (RBFN)을 이용하여 플라즈마 전자밀도를 모델링하였다. RBFN의 예측성능은 학습인자의 함수로 최적화하였다. 체계적인 모델링을 위해 통계적인 실험계획법이 적용되었으며, 실험은 반구형 유도결합형 플라즈마 장비를 이용하여 수행이 되었다. 전자밀도측정에는 Langmuir probe가 이용되었다. 최적화된 RBFN모델을 통계적인 회귀 모델과 비교하였으며, 59%정도 모델의 예측성능을 향상시켰다.

• 한국음향학회:학술대회논문집
• /
• 한국음향학회 1999년도 학술발표대회 논문집 제18권 2호
• /
• pp.71-76
• /
• 1999

본 논문에서 제안한 예측형 RBFN(Radial Basis Function Network)은 HMM과 신경망을 결합한 하이브리드 구조이다. 이 신경망은 HMM으로 추정한 확률분포 파라미터를 사용하여 중간층의 활성화 함수의 출력을 결정하고, 중간층과 출력층의 연결강도만 네트워크 내에서 학습한다. 그리고 HMM으로 추정한 확률분포 파라미터는 두 가지 방법으로 예측형 RBFN에 이용하였다. 첫 번째는 HMM의 각 상태의 혼합수 만큼의 중간층 유니트를 주는 방법이고, 두 번째는 HMM의 혼합수$\times$출력분포수 만큼의 중간층 유니트를 주는 방법이다. 실험결과, 예측형 RBFN은 다른 방법들의 결과보다 $4.5\~6.5\%$ 저하된 결과를 보였지만 다른 신경망에 비해서 학습 반복 횟수를 작게할 수 있었으며 전체 학습시간을 대폭 단축할 수 있었다.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2001년도 ICCAS
• /
• pp.26.6-26
• /
• 2001

In this paper we introduce a new algorithm that enables radial basis function network(RBFN) to be structured automatically and guarantees the stability of the RBFN. Because this new algorithm is efficient and also have the advantage of fast computational speed we adopt this algorithm as online learning scheme for uncertain nonlinear dynamical systems. Based on the fact that a 3-layered RBFN can represent a specific nonlinear function reasonably well by linearly combining a set of nonlinear and localized basis functions, we show that this RBFN can identify the nonlinear system very well without knowing the information of the system in advance.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1999년도 하계학술대회 논문집 G
• /
• pp.2992-2994
• /
• 1999

This paper investigates the direct adaptive control of nonlinear systems using RBFN(radial basis function networks). The structure of the controller consists of a fixed PD controller and a RBFN controller in parallel. An adaptation law for the weight adjustment is developed based on the Lyapunov stability theory to guarantee the stability of the overall control scheme. Also, the tracking errors between the system outputs and the desired outputs converge to zero asymptotically. To evaluate the performance of the controller, the proposed method is applied to the trajectory control of the two-link manipulator.

• 전기학회논문지P
• /
• 제57권4호
• /
• pp.362-369
• /
• 2008

In this study, we developed RBFN(Radial Basis Function Networks) based load modeling method with harmonic components. The developed method considers harmonic information as well as fundamental frequency and voltage considered as essential factors in conventional method. Thus, the proposed method makes it possible to effectively estimate load characteristics in power lines with harmonics. RBFN has some advantage such as simple structure and rapid computation ability compared with multi-layer perceptorn which is extensively applied for load modeling. To verify the effectiveness, the proposed method has been intensively tested with various dataset acquired under the different frequency and voltage and compared it with conventional methods such as polynomial method, MLPN and RBFN with no harmonic components.

• 한국지능시스템학회:학술대회논문집
• /
• 한국퍼지및지능시스템학회 1997년도 추계학술대회 학술발표 논문집
• /
• pp.311-314
• /
• 1997

This paper presents recognition of high impedance fault patterns based of chaotic features using the Radial Basis Function Network(RBFN). The chaos attractor is reconstructed from the fault current data for pattern recognition. The RBFN successfully classifies the three kinds of fault pattems and one normal pattem.

• 한국지능시스템학회논문지
• /
• 제8권6호
• /
• pp.122-128
• /
• 1998

이 논문에서는 영상시퀀스의 프레임간 차영상 블록을 영상활동도의 크기 및 분포에 따라 적응적으로 분류함으로써 영상시퀀스를 압축하는 기법을 제안한다. 활동도의 크기에 의한 분류에서는 차영상 블록에 포함되어 있는 물체의 에지부분에 해당하는 활동블록과 비활동 블록으로 분류하고, 활동도의 분포에 의한 분류에서도 활동블록들을 이산 코사인변환계수의 분포정도를 특징으로 하여 수직, 수평, 저활동 블록으로 분류한다. 대표적인 분류결과를 이용하여 RBFN 신경망을 학습시켜 프레임간 차영상 블록들의 비선형적인 분류 특성을 얻었다. 시뮬레이션 결과 RBFN을 이용한 차영상 블록의 분류가 영상활동도의 정렬방법이나 다층퍼셉트론 신경망(MLP)에 비해 영상시퀀스의 압축성능이 향상되었다.

• 한국전기전자재료학회논문지
• /
• 제18권1호
• /
• pp.1-5
• /
• 2005

A new model of plasma etch process was constructed by using a radial basis function network (RBFN). This technique was applied to an etching of silicon carbide films in a NF$_3$ inductively coupled plasma. Experimental data to train RBFN were systematically collected by means of a 2$^4$ full factorial experiment. Appropriateness of prediction models was tested with test data consisted of 16 experiments not pertaining to the training data. Prediction performance was optimized with variations in three training factors, the number of pattern units, width of radial basis function, and initial weight distribution between the pattern and output layers. The etch responses to model were an etch rate and a surface roughness measured by atomic force microscopy. Optimized models had the root mean-squared errors of 26.1 nm/min and 0.103 nm for the etch rate and surface roughness, respectively. Compared to statistical regression models, RBFN models demonstrated an improvement of more than 20 % and 50 % for the etch rate and surface roughness, respectively. It is therefore expected that RBFN can be effectively used to construct prediction models of plasma processes.