• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.35S no.6
• /
• pp.46-54
• /
• 1998

This paper presents a method for designing robust fuzzy $H^{\infty}$ controllers which stabilize nonlinear systems with parameter uncertainty adn guarantee an induced $L_{2}$ norm bound constraint on disturbance attenuation for all admissible uncertainties. Takagi and Sugeno's fuzzy models with uncertainty are used as the model for the uncertain nonlinear systems. Fuzzy control systems utilize the concept of so-called parallel distributed compensation(PDC). Using a single quadratic Lyapunov function, the stability condition satisfying decay rate and disturbance attenuation condition for Takagi and Sugeno's fuzzy model with parameter uncertainty are discussed. A sufficient condition for the existence of robust fuzzy $H^{\infty}$ controllers is then presented in terms of linear matrix inequalities(LMIs). Finally, design examples of robust fuzzy $H^{\infty}$ controllers for uncertain nonlinear systems are presented.

• IEIE Transactions on Smart Processing and Computing
• /
• v.2 no.5
• /
• pp.277-281
• /
• 2013

Modern speaker verification systems based on support vector machines (SVMs) use Gaussian mixture model (GMM) supervectors as their input feature vectors, and the maximum a posteriori (MAP) adaptation is a conventional method for generating speaker-dependent GMMs by adapting a universal background model (UBM). MAP adaptation requires the appropriate amount of input utterance due to the number of model parameters to be estimated. On the other hand, with limited utterances, unreliable MAP adaptation can be performed, which causes adaptation noise even though the Bayesian priors used in the MAP adaptation smooth the movements between the UBM and speaker dependent GMMs. This paper proposes a sparse MAP adaptation method, which is known to perform well in the automatic speech recognition area. By introducing sparse MAP adaptation to the GMM-SVM-based speaker verification system, the adaptation noise can be mitigated effectively. The proposed method utilizes the L0 norm as a regularizer to induce sparsity. The experimental results on the TIMIT database showed that the sparse MAP-based GMM-SVM speaker verification system yields a 42.6% relative reduction in the equal error rate with few additional computations.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.6
• /
• pp.2806-2825
• /
• 2018

Unsupervised learning has shown good performance on image, video and audio classification tasks, and much progress has been made so far. It studies how systems can learn to represent particular input patterns in a way that reflects the statistical structure of the overall collection of input patterns. Many promising deep learning systems are commonly trained by the greedy layerwise unsupervised learning manner. The performance of these deep learning architectures benefits from the unsupervised learning ability to disentangling the abstractions and picking out the useful features. However, the existing unsupervised learning algorithms are often difficult to train partly because of the requirement of extensive hyperparameters. The tuning of these hyperparameters is a laborious task that requires expert knowledge, rules of thumb or extensive search. In this paper, we propose a simple and effective unsupervised feature learning algorithm for image classification, which exploits an explicit optimizing way for population and lifetime sparsity. Firstly, a sparse target matrix is built by the competitive rules. Then, the sparse features are optimized by means of minimizing the Euclidean norm ($L_2$) error between the sparse target and the competitive layer outputs. Finally, a classifier is trained using the obtained sparse features. Experimental results show that the proposed method achieves good performance for image classification, and provides discriminative features that generalize well.

• International Journal of CAD/CAM
• /
• v.6 no.1
• /
• pp.59-64
• /
• 2006

A new method to obtain explicit re-parameterization that preserves the curve degree and parametric domain is presented in this paper. The re-parameterization brings a curve very close to the arc length parameterization under $L_2$ norm but with less segmentation. The re-parameterization functions we used are $C^1$ continuous piecewise rational linear functions, which provide more flexibility and can be easily identified by solving a quadratic equation. Based on the outstanding performance of Mobius transformation on modifying pieces with monotonic parametric speed, we first create a partition of the original curve, in which the parametric speed of each segment is of monotonic variation. The values of new parameters corresponding to the subdivision points are specified a priori as the ratio of its cumulative arc length and its total arc length. $C^1$ continuity conditions are imposed to each segment, thus, with respect to the new parameters, the objective function is linear and admits a closed-form optimization. Illustrative examples are also given to assess the performance of our new method.

• Journal of Korean Society of Forest Science
• /
• v.86 no.3
• /
• pp.301-310
• /
• 1997

Users' norms, expectations and recreation motives are major influential factors on perceived conflict in recreation settings. This study was conducted to examine how information affects users' norms and expectations, and subsequently, the extent of perceived conflict. Data was collected at the Second Campground in Chirisan National Park in 1994. Of two hundred eighty on-site survey questionnaires distributed, 253(90.4%) were used for analysis. According to the result of the study, information changed users' normative standards in accord with the content of information and increased the extent of norm convergence. However, it did not change users' expectations and didn't reduce the degree of perceived canflict. The possible reasons were discussed. Some directions for more effective use of information to reduce perceived conflict were also suggested.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.9
• /
• pp.4510-4526
• /
• 2017

Most available methods of facial gender recognition work well under a constrained situation, but the performances of these methods have decreased significantly when they are implemented under unconstrained environments. In this paper, a method via low-rank and collaborative representation is proposed for facial gender recognition in the wild. Firstly, the low-rank decomposition is applied to the face image to minimize the negative effect caused by various corruptions and dynamical illuminations in an unconstrained environment. And, we employ the collaborative representation to be as the classifier, which using the much weaker $l_2-norm$ sparsity constraint to achieve similar classification results but with significantly lower complexity. The proposed method combines the low-rank and collaborative representation to an organic whole to solve the task of facial gender recognition under unconstrained environments. Extensive experiments on three benchmarks including AR, CAS-PERL and YouTube are conducted to show the effectiveness of the proposed method. Compared with several state-of-the-art algorithms, our method has overwhelming superiority in the aspects of accuracy and robustness.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.28 no.5
• /
• pp.503-510
• /
• 2004

In this paper, the well-conditioned observer for a stochastic system is designed so that the observer is less sensitive to the ill-conditioning factors in transient and steady-state observer performance. These factors include not only deterministic uncertainties such as unknown initial estimation error, round-off error, modeling error and sensing bias, but also stochastic uncertainties such as disturbance and sensor noise. In deterministic perspectives, a small value in the L$_{2}$ norm condition number of the observer eigenvector matrix guarantees robust estimation performance to the deterministic uncertainties. In stochastic viewpoints, the estimation variance represents the robustness to the stochastic uncertainties and its upper bound can be minimized by reducing the observer gain and increasing the decay rate. Both deterministic and stochastic issues are considered as a weighted sum with a LMI (Linear Matrix Inequality) formulation. The gain in the well-conditioned observer is optimally chosen by the optimization technique. Simulation examples are given to evaluate the estimation performance of the proposed observer.

• Journal of Nutrition and Health
• /
• v.47 no.6
• /
• pp.385-393
• /
• 2014

Purpose: The aim of this study is to examine the effects of egg consumption and suggest proper guidelines for consumption of eggs by determining the relationship between eggs and cholesterol. Methods: Literature review was conducted on the relationship between nutritional, functional properties of eggs and serum cholesterol, as well as cardiovascular disease. Results: Eggs, which are a good protein food with complete amino acid composition, contain vitamin A, riboflavin, vitamin $B1_2$, folic acid, vitamin D, vitamin E, vitamin K, calcium, iron, choline, selenium, ${\beta}$-carotene, lutein, zeaxanthin, etc. However the egg yolk has a high cholesterol content, which is associated with chronic diseases, including heart disease and hypertension. As a result, its intake is subject to regulation. Outbreak of heart disease by yolk intake can show different results depending on the characteristics of the subjects, amount of egg intake, and the implications of other foods eaten. It is difficult to determine whether eggs are beneficial, as they are the main supplying source for other major nutritive elements as well. Several research studies insist that when cholesterol intake increases by 100 mg, the level of serum cholesterol increases by 2.2~4.5 mg/dL and when serum cholesterol increases by 1%, the risk of heart disease increases by 2%. This indicates that a large intake of eggs can increase the risk of heart disease. Although the cholesterol of egg yolk and serum cholesterol are correlated, it is insufficient to conclude that only cholesterol and not other components are related to heart disease. In fact, other components in egg such as various unsaturated fatty acids and phospholipids could be related as well. Rather than concluding egg as a 'good' or 'bad' food according to its cholesterol content, it is important to define egg as a part of dietary patterns. Conclusion: Generalizing an indiscriminate and uniform amount of egg intake for all seems inadequate. However, patients with diabetes or heart disease should pay particular attention to the amount of egg intake. As for the norm, eating egg with vegetables as a substitute for other animal products seems beneficial.

• Journal of Biomedical Engineering Research
• /
• v.24 no.1
• /
• pp.1-8
• /
• 2003

The job of long distance driving is likely to be fatiguing and requires long period alertness and attention, which make considerable demands of the driver. Driving fatigue contributes to driver related with accidents and fatalities. In this study, we investigated the relationship between the number of hours of driving and driving fatigue using heart rate variability(HRV) signal. With a more traditional measure of overall variability (standard deviation, mean, spectral values of heart rate). Nonlinear characteristics of HRV signal were analyzed using Approximate Entropy (ApEn) and Poincare plot. Five subjects drive the four passenger vehicle twice. All experiment number was 40. The test route was about 300Km continuous long highway circuit and driving time was about 3 hours. During the driving, measures of electrocardiogram(ECG) were performed at intervals of 30min. HRV signal, derived from the ECG, was analyzed using time, frequency domain parameters and nonlinear characteristic. The significance of differences on the response to driving fatigue was determined by Student's t-test. Differences were considered significant when a p value < 0.05 was observed. In the results, mean heart rate(HRmean) decreased consistently with driving time, standard deviation of RR intervals(SDRR), standard deviation of the successive difference of the RR intervals(SDSD) increased until 90min. Hereafter, they were almost unchanging until the end of the test. Normalized low frequency component $(LF_{norm})$, ratio of low to high frequency component (LF/HF) increased. We used the Approximate Entropy(ApEn), Poincare plot method to describe the nonlinear characteristics of HRV signal. Nonlinear characteristics of HRV signals decreased with driving time. Statistical significant is appeared after 60 min in all parameters.

• Steel and Composite Structures
• /
• v.5 no.2_3
• /
• pp.181-192
• /
• 2005

The effective length method for flexural (column) buckling has been used for many decades but its use is somewhat limited in various contemporary design codes to moderately slender structures with elastic critical load factor (${\lambda}_{cr}$) less than 3 to 5. In pace with the use of higher grade steel in recent years, the influence of buckling in axial buckling resistance of a column becomes more important and the over-simplified assumption of effective length factor can lead to an unsafe, an uneconomical or a both unsafe and uneconomical solution when some members are over-designed while key elements are under-designed. Effective length should not normally be taken as the distance between nodes multiplied by an arbitrary factor like 0.85, 1.0, 2.0 etc. Further, the classification of non-sway and sway-sensitive frames makes the conventional design procedure tedious to use and, more importantly, limited to simple regular frames. This paper describes the practical use of second-order analysis with section capacity check allowing for $P-{\delta}$ and $P-{\Delta}$ effects together with member and system imperfections. Most commercial software considers only the $P-{\Delta}$ effect, but not member and frame imperfections nor $P-{\delta}$ effect, and engineers must be very careful in their uses. A verification problem is also given for validation of software for this type of powerful second-order analysis and design. It is a trend for popular and advanced national design codes in using the second-order analysis as a norm for analysis and design of steel structures while linear analysis may only be used in very simple structures.