• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 2000년도 추계학술대회 논문집
• /
• pp.687-690
• /
• 2000

Lately, many studies have been progressed for the protection human's lives and property as holding in check accidents happened by human's carelessness or mistakes. One part of these is the development of an autonomouse vehicle. General control method of vision-based autonomous vehicle system is to determine the navigation direction by analyzing lane images from a camera, and to navigate using proper control algorithm. In this paper, characteristic points are abstracted from lane images using lane recognition algorithm with sobel operator. And then the vehicle is controlled using two proposed auto-steering algorithms. Two steering control algorithms are introduced in this paper. First method is to use the geometric relation of a camera. After transforming from an image coordinate to a vehicle coordinate, a steering angle is calculated using Ackermann angle. Second one is using a neural network algorithm. It doesn't need to use the geometric relation of a camera and is easy to apply a steering algorithm. In addition, It is a nearest algorithm for the driving style of human driver. Proposed controller is a multilayer neural network using Levenberg-Marquardt backpropagation learning algorithm which was estimated much better than other methods, i.e. Conjugate Gradient or Gradient Decent ones.

• International Journal of Control, Automation, and Systems
• /
• 제1권2호
• /
• pp.210-214
• /
• 2003

This paper presents a residence time distribution (RTD) measurement method using a radioisotope tracer and the estimation method of RTD model parameters to analyze a submerged bio-reactor. The mathematical RTD models have been investigated to represent the flow behavior and the existence of stagnant regions in the reactor. Knowing the parameters of the RTD model is important for understanding the mixing characteristics of a reactor The radioisotope tracer experiment was carried out by injecting a radioisotope tracer as a pulse into the inlet of the reactor and recording the change of its concentration at the outlet of the reactor to obtain the experimental RTD response. The parameter estimation was performed by the Levenberg-Marquardt optimization algorithm. The proposed scheme allowed the parameter estimation of RTD model suggested by Adler-Hovorka with very low deviations. The estimation procedure is shown to lead to accurate estimation of the RTD parameters and to a good agreement between experimental and simulated response.

• 대한전자공학회논문지SP
• /
• 제38권6호
• /
• pp.657-664
• /
• 2001

최근에 물체영상들로부터 3차원 물체 모델을 복원할 수 있는 셀프캘리브레이션 기술에 대한 연구가 활발히 진행되고 있다. 이 셀프캘리브레이션 기술의 핵심은 F행렬이며, 복원되는 3차원 물체 모델의 정확도는 물체영상들 사이에서 유도해내는 F행렬의 추정의 정확도에 좌우된다. F행렬을 추정하기 위해 일반적으로 선형최소화방법이 적용되고있다. 그러나 본 논문에서는 보다더 정확한 F행렬의 추정을 위해 비선형 최소화방법인 Levenberg-Marquardt 기법을 적용하였다. 또한 F행렬의 정확도를 감소시키는 부정확한 대응점들 (corresponding points)과 오차를 많이 포함하고 있는 대응점들, 즉 outliers를 Monte Carlo 기술을 적용하여 제거하였다. 본 논문에서 적용한 방법들로 추정한 F행렬의 정확도를 분석한 결과, outliers를 제거하기 전보다 제거한 후의 정확도가 31% 향상되었고, 선형적 추정 F행렬보다 비선형적 추정 F행렬이 22% 향상되었음을 알 수 있었다.

• 산업기술연구
• /
• 제19권
• /
• pp.127-133
• /
• 1999

This paper is a research about system identification which optimizes uncertain geothechnical properties from the data measured during geotechnical design and construction. Various numerical optimization algorithms of Simplex method, Powell method, Rosenbrock method and Levenberg-Marquardt method were applied to the excavation problem to determine which method showed the best results with respect to robustness of success in finding an optimal solution to within a certain accuracy and number of function evaluations. From the results of numerical analysis, all of four algorithms are converged to exact solution after satisfying the allowed criteria, and Levenberg-Marquardt's algorithms was identified to be the most efficient method in number of function evaluations. System identification was applied to geotechnical engineering problems, possibly being occurred in field, to verify its applicability : estimation of settlement due to self-weight consolidation in dredged and filled soil. For self-weight consolidational settlement of a dredged soil, a program of evaluating the constitutive relationship of effective stress-void ratio-permeability was developed by using the technique of system identification. Thus, consolidational characteristics of a dredged soil, having a very high initial void ratio, can be evaluated.

• 대한의용생체공학회:의공학회지
• /
• 제21권2호
• /
• pp.157-163
• /
• 2000

최근 MR영상을 허혈성 뇌졸중의 초급성기에 뇌조직의 관류 이상을 조기에 진단하려는 연구들이 진행되고 있으나 아직 일반적인 진단용 소프트웨어만 있을 뿐 영상 자료를 후처리하여 뇌조직의 구조 및 기능적인 정보를 제공하는 mapping 영상을 특수 소프트웨어는 실용화되어 있지 않다. 본 논문에서는 Gamma-variate 곡선 정합을 이용한 뇌관류 파라미터 영상 mapping의 알고리즘 구현에 관해 연구하였다. 관류 MR영상의 각 화소마다 측정된 시간에 따른 신호강도의 변화 곡선은 비선형적이어서 뇌관류에 관한 여러 가지 혈역학적 변수들을 보다 정확하게 계산할 수 없었다. 그래서 수렴속도가 빠르고 안정성이 높은 비선형 최적화 알고리즘인 Levenberg-Marquardt 알고리즘(LMA)을 활용하였다. 즉 시간에 따른 신호강도의 변화 곡선을 Gamma-variate 함수를 이용하여 곡선 정합한 후, CBV, MTT, CBF, TTP, BAT, MS의 여러 가지 혈역학적 변수를 LMA에 의해 계산하였다. 그 결과로 관류 MR영상으로부터 얻은 mapping 영상은 초급성 허혈성 뇌졸중에서 관류에 관한 혈역학적 변화를 평가함으로써 나중에 생길 뇌경색의 범위를 예견하는 데에 유용하였다.

• 한국추진공학회지
• /
• 제11권5호
• /
• pp.1-13
• /
• 2007

미래형 추진기관으로 주목받고 있는 스크램제트 엔진의 지상시험을 위해 시험모델을 설계하였다. 설계 마하수는 7.6, 고도는 30km로 두었으며 4개의 충격파를 흡입구에 배치하였다. 엔진의 흡입구는 Levenberg-Marquardt 최적화 기법, Korkegi 관계식을 이용하여 설계하였으며 연소기는 연료-공기 혼합 증진을 통하여 고연소효율 및 연소기 길이 단축을 구현할 수 있도록 설계하였다. 성능검증을 위한 전산해석에서 흡입구는 받음각 ${\pm}4^{\circ}$에서도 적절한 충격파배치를 보였으며 연소기는 공동을 설치하였을 때 연소효율이 향상됨을 확인할 수 있었다.

• 한국항해항만학회:학술대회논문집
• /
• 한국항해항만학회 2006년도 International Symposium on GPS/GNSS Vol.2
• /
• pp.101-106
• /
• 2006

The geoidal undulations are needed for determining the orthometric heights from the Global Positioning System GPS-derived ellipsoidal heights. There are several methods for geoidal undulation determination. The paper presents a method employing a simple architecture Two Layer Multiquadric-Biharmonic Artificial Neural Network (TLMB-ANN) to approximate an area of 4200 square kilometres quasigeoid surface with GPS-levelling data. Hardy’s Multiquadric-Biharmonic functions is used as the hidden layer neurons’ activation function and Levenberg-Marquardt algorithm is used to train the artificial neural network. In numerical examples five surfaces were compared: the gravimetric geometry hybrid quasigeoid, Support Vector Machine (SVM) model, Hybrid Fuzzy Neural Network (HFNN) model, Traditional Three Layer Artificial Neural Network (ANN) with tanh activation function and TLMB-ANN surface approximation. The effectiveness of TLMB-ANN surface approximation depends on the number of control points. If the number of well-distributed control points is sufficiently large, the results are similar with those obtained by gravity and geometry hybrid method. Importantly, TLMB-ANN surface approximation model possesses good extrapolation performance with high precision.

• Journal of Electrical Engineering and Technology
• /
• 제9권3호
• /
• pp.1060-1071
• /
• 2014

Signals of the Electroencephalogram (EEG) can reflect the electrical background activity of the brain generated by the cerebral cortex nerve cells. This has been the mostly utilized signal, which helps in effective analysis of brain functions by supervised learning methods. In this paper, an approach for improving the accuracy of EEG signal classification is presented to detect epileptic seizures. Moreover, Independent Component Analysis (ICA) is incorporated as a preprocessing step and Short Time Fourier Transform (STFT) is used for denoising the signal adequately. Feature extraction of EEG signals is accomplished on the basis of three parameters namely, Standard Deviation, Correlation Dimension and Lyapunov Exponents. The Artificial Neural Network (ANN) is trained by incorporating Levenberg-Marquardt(LM) training algorithm into the backpropagation algorithm that results in high classification accuracy. Experimental results reveal that the methodology will improve the clinical service of the EEG recording and also provide better decision making in epileptic seizure detection than the existing techniques. The proposed EEG signal classification using feed forward Backpropagation Neural Network performs better than to the EEG signal classification using Adaptive Neuro Fuzzy Inference System (ANFIS) classifier in terms of accuracy, sensitivity, and specificity.

• 한국항공우주학회지
• /
• 제47권7호
• /
• pp.499-507
• /
• 2019

고정익 항공기의 수학적 모델이 잘 알려져 있음에도 불구하고, 넓은 비행 영역에서 모델링 오차를 고려하여 설계 제어 성능을 달성하기 위한 다양한 연구가 있다. 본 논문은 레벤버그-마쿼트 알고리듬을 적용한 모델-참조 적응 제어 법칙과, 이를 이용한 고정익 무인항공기의 피치 자세 제어에 대한 연구를 소개한다. 또한 모델-참조 적응 제어의 기준 모델을 모델의 동특성에 기인하여 결정함으로써 성능지표를 제시한다. 설계한 적응 법칙의 성능은 시뮬레이션과 비행실험을 통해 검증했다.

• Nuclear Engineering and Technology
• /
• 제53권8호
• /
• pp.2708-2715
• /
• 2021

The Multi-Purpose Overload Robot (CMOR) is a key subsystem of China Fusion Engineering Test Reactor (CFETR) remote handling system. Due to the long cantilever and large loads of the CMOR, it has a large rigid-flexible coupling deformation that results in a poor position accuracy of the end-effector. In this study, based on the Levenberg-Marquardt algorithm, the spatial grid, and the linearized variable load principle, a variable parameter compensation model was designed to identify the parameters of the CMOR's kinematics models under different loads and at different poses so as to improve the trajectory tracking accuracy. Finally, through Adams-MATLAB/Simulink, the trajectory tracking accuracy of the CMOR's rigid-flexible coupling model was analyzed, and the end position error exceeded 0.1 m. After the variable parameter compensation model, the average position error of the end-effector became less than 0.02 m, which provides a reference for CMOR error compensation.