• Proceedings of the Korean Information Science Society Conference
• /
• 2004.04b
• /
• pp.721-723
• /
• 2004

A trust-region method is a quite attractive optimization technique. It is, in general, faster than the steepest descent method and is free of a learning rate unlike the gradient-based methods. In addition to its convergence property (between linear and quadratic convergence), ifs stability is always guaranteed, in contrast to the Newton's method. In this paper, we present an efficient implementation of the maximum likelihood independent component analysis (ICA) using the trust-region method, which leads to trust-region-based ICA (TR-ICA) algorithms. The useful behavior of our TR-ICA algorithms is confimed through numerical experimental results.

• International journal of advanced smart convergence
• /
• v.8 no.2
• /
• pp.39-46
• /
• 2019

Human mobility estimation plays a key factor in a lot of promising applications including location-based recommendation systems, urban planning, and disease outbreak control. We study the human mobility estimation problem in the case where recent locations of a person-of-interest are unknown. Since matrix decomposition is used to perform latent semantic analysis of multi-dimensional data, we propose a human location estimation algorithm based on matrix factorization to reconstruct the human movement patterns through the use of information of persons with correlated movements. Specifically, the optimization problem which minimizes the difference between the reconstructed and actual movement data is first formulated. Then, the gradient descent algorithm is applied to adjust parameters which contribute to reconstructed mobility data. The experiment results show that the proposed framework can be used for the prediction of human location and achieves higher predictive accuracy than a baseline model.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.57 no.6
• /
• pp.915-923
• /
• 2008

This paper focuses on optimal tuning of nonlinear parameters of a dual-input power system stabilizer(dual-input PSS), which can improve the system damping performance immediately following a large disturbance. Until recently, various PSS models have developed to bring stability and reliability to power systems, and some of these models are used in industry applications. However, due to non-smooth nonlinearities from the interaction between linear parameters(gains and time constants of linear controllers) and nonlinear parameters(saturation output limits), the output limit parameters cannot be determined by the conventional tuning methods based on linear analysis. Only ad hoc tuning procedures('trial and error' approach) have been used. Therefore, the steepest descent method is applied to implement the optimal tuning of the nonlinear parameters of the dual-input PSS. The gradient required in this optimization technique can be computed from trajectory sensitivities in hybrid system modeling with the differential-algebraic-impulsive-switched(DAIS) structure. The optimal output limits of the dual-input PSS are evaluated by time-domain simulation in both a single machine infinite bus(SMIB) system and a multi-machine power system in comparison with those of a single-input PSS.

• Journal of Institute of Control, Robotics and Systems
• /
• v.15 no.12
• /
• pp.1175-1183
• /
• 2009

In photovoltaic(PV) generator systems, DC-DC converters are significantly considered for control system performance in power quality point of view. This paper presents a novel adaptive control method for DC-DC converters applied in PV generator systems. First, we derive a state-space average model of the converter system and then propose a reset control methodology to enhance transient response performance for time-varying PV systems. For estimating parameters of a reset control, a gradient descent optimization is utilized and an adjustment rule of them are derived respectively. An objective of the optimization is that characteristic equation of an augmented system model which is formed with an converter system model and an reset control is to trace a predefined polynomial given as a reference characteristic model. Next, we accomplish stability analysis by means of a well-known Lyapunov theory for nonlinear converter systems including time-varying voltage excitation from a PV generator. Numerical simulation demonstrates reliability of our control methodology and its superiority by comparison to a traditional control strategy.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.14 no.6
• /
• pp.965-973
• /
• 2011

Small unmanned ground vehicles(SUGVs) are typically operational on unstructured environments such as crashed building, mountain area, caves, and so on. On those terrains, driving control can suffer from the unexpected ground disturbances which occasionally lead turnover situation. In this paper, we have proposed an algorithm which sustains driving stability of a SUGV as preventing from turnover. The algorithm exploits potential field method in order to determine the stability of the robot. Then, the flipper and manipulator posture of the SUGV is optimized from local optimization algorithm known as gradient descent method. The proposed algorithm is verified using 3D dynamic simulation, and results showed that the proposed algorithm contributes to driving stability of SUGV.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.16 no.3
• /
• pp.81-87
• /
• 2016

This paper analysis the AM-SCS-MMA (Adaptive Modulus-Soft Constraint Satisfaction-MMA) based on the adaptive modulus and minimus-disturbance technique in order to improve the stability and robustness in low signal to noise power of current MMA adaptive equalization algorithm. In AM-SCS-MMA, it updates the filter coefficient applying the adaptive modulus and minimum-disturbance technique of deterministic optimization problem instead of LMS or gradient descend algorithm for obtain the minimize the cost function of adaptive equalization. It is possible to improve the equalizer filter stability, robustness to the various noise characteristic and simultaneous reducing the intersymbol interference due to the amplitude and phase distortion occurred at channel. The computer simulation were performed for confirming the improved performance of SCS-MMA. For these, the output signal constellation of equalizer, residual isi, MSE, EMSE (Excess MSE) which means the channel traking capability and SER which means the robustness were applied. As a result of computer simulation, the AM-SCS-MMA have slow convergence time and less residual quantities after steady state, more good robustness in the poor signal to noise ratio, but poor in channel tracking capabilities was confirmed than MMA.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.7
• /
• pp.1-7
• /
• 2021

In this paper, we propose a feature selection technique for multi-label classification. Many existing feature selection techniques have selected features by calculating the relation between features and labels such as a mutual information scale. However, since the mutual information measure requires a joint probability, it is difficult to calculate the joint probability from an actual premise feature set. Therefore, it has the disadvantage that only a few features can be calculated and only local optimization is possible. Away from this regional optimization problem, we propose a feature selection technique that constructs a low-rank space in the entire given feature space and selects features with sparsity. To this end, we designed a regression-based objective function using Nuclear norm, and proposed an algorithm of gradient descent method to solve the optimization problem of this objective function. Based on the results of multi-label classification experiments on four data and three multi-label classification performance, the proposed methodology showed better performance than the existing feature selection technique. In addition, it was showed by experimental results that the performance change is insensitive even to the parameter value change of the proposed objective function.

• Korean Journal of Remote Sensing
• /
• v.33 no.4
• /
• pp.423-436
• /
• 2017

This paper presents a deep learning-based road segmentation framework from very high-resolution orthophotos. The proposed method uses Deep Convolutional Autoencoders for end-to-end mapping of orthophotos to road segmentations. In addition, a set of post-processing steps were applied to make the model outputs GIS-ready data that could be useful for various applications. The optimization of the model's parameters is explained which was conducted via grid search method. The model was trained and implemented in Keras, a high-level deep learning framework run on top of Tensorflow. The results show that the proposed model with the best-obtained hyperparameters could segment road objects from orthophotos at an average accuracy of 88.5%. The results of optimization revealed that the best optimization algorithm and activation function for the studied task are Stochastic Gradient Descent (SGD) and Exponential Linear Unit (ELU), respectively. In addition, the best numbers of convolutional filters were found to be 8 for the first and second layers and 128 for the third and fourth layers of the proposed network architecture. Moreover, the analysis on the time complexity of the model showed that the model could be trained in 4 hours and 50 minutes on 1024 high-resolution images of size $106{\times}106pixels$, and segment road objects from similar size and resolution images in around 14 minutes. The results show that the deep learning models such as Convolutional Autoencoders could be a best alternative to traditional machine learning models for road segmentation from aerial photographs.

• Proceedings of the Korea Electromagnetic Engineering Society Conference
• /
• 2005.11a
• /
• pp.169-174
• /
• 2005

In this paper, we proposed a new algorithm of the inverse scattering for the reconstruction of unknown dielectric scatterers using the finite-difference time-domain method and the design sensitivity analysis. We introduced the design sensitivity analysis based on the gradient for the fast convergence of the reconstruction. By introducing the adjoint variable method for the efficient calculation, we derived the adjoint variable equation. As an optimal algorithm we used the steepest descent method and reconstructed the dielectric targets using the iterative estimation. To verify our algorithm we will show the numerical examples for the two-dimensional $TM^2$ cases.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.25 no.12
• /
• pp.2062-2069
• /
• 2001

It has been established that a crack has an important effect on the dynamic behavior of a structure. This effect depends mainly on the location and depth of the crack. To identify the location and depth of a crack in a structure, a method is presented in this paper which uses synthetic artificial intelligent technique, that is, Adaptive-Network-based Fuzzy Inference System(ANFIS) solved via hybrid learning algorithm(the back-propagation gradient descent and the least-squares method) are used to learn the input(the location and depth of a crack)-output(the structural eigenfrequencies) relation of the structural system. With this ANFIS and a continuous evolutionary algorithm(CEA), it is possible to formulate the inverse problem. CEAs based on genetic algorithms work efficiently for continuous search space optimization problems like a parameter identification problem. With this ANFIS, CEAs are used to identify the crack location and depth minimizing the difference from the measured frequencies. We have tried this new idea on a simple beam structure and the results are promising.