• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.25 no.7A
• /
• pp.967-977
• /
• 2000

In this paper, we apply a mixed $H_2/H_{\infty}$ control to a generalized plant of inverted pendulum system represented by an LFR(Linear Fractional Representation). First, in order to obtain the generalized plant, the linear model of the inverted pendulum represented by an LFR(Linear fractional Representation) is derived. In LFR, we consider system uncertainties as three nonlinear components and a pendulum mass uncertainty. Augmenting the LFR model by adding weighting functions, we get a generalized plant. And then, we design a mixed $H_2/H_{\infty}$ controller for the generalized plant. In order to design the mixed $H_2/H_{\infty}$ controller, we use the LMI technique. To evaluate control performances and robust stability of the mixed $H_2/H_{\infty}$ controller designed, we compare it with the $H_{\infty}$ controller through the simulation and experiment. In the result, with the fewer feedback information, the mixed $H_2/H_{\infty}$ controller shows the better control performances and robust stability than the $H_{\infty}$ controller in the sense of pendulum angle.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.41 no.10
• /
• pp.931-940
• /
• 2017

This paper describes a predictive fault diagnosis algorithm for autonomous vehicles based on a kinematic model that uses a sliding mode observer. To ensure the safety of autonomous vehicles, reliable information about the environment and vehicle dynamic states is required. A predictive algorithm that can interactively diagnose longitudinal environment and vehicle acceleration information is proposed in this paper to evaluate the reliability of sensors. To design the diagnosis algorithm, a longitudinal kinematic model is used based on a sliding mode observer. The reliability of the fault diagnosis algorithm can be ensured because the sliding mode observer utilized can reconstruct the relative acceleration despite faulty signals in the longitudinal environment information. Actual data based performance evaluations are conducted with various fault conditions for a reasonable performance evaluation of the predictive fault diagnosis algorithm presented in this paper. The evaluation results show that the proposed diagnosis algorithm can reasonably diagnose the faults in the longitudinal environment and acceleration information for all fault conditions.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.11 no.9
• /
• pp.3304-3310
• /
• 2010

This study solves the decision making problems for TFT-LCD manufacturing supply chain with demand and price uncertainties by establishing robust production and distribution strategies. In order to control the decisions regarding production graded by quality, inventory level and distribution, this study develop scenario model based stochastic mixed integer linear programs (SMILPs) that consider demand and price uncertainties as well as realistic constraints such as capacities etc. The performance of the solution obtained from the SMILPs using robust algorithms will be evaluated through various scenarios.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.42 no.1
• /
• pp.79-87
• /
• 2005

This paper presents an active contour model for extracting accurate facial regions in complex environments. In the model, a contour is represented by a zero level set of level function φ, and evolved via level set partial differential equations. Then, unlike general active contours, skin color information that is represented by 2D Gaussian model is used for evolving and slopping a curve, which allows the proposed method to be robust to noise and varying pose. To assess the effectiveness of the proposed method it was tested with several natural scenes, and the results were compared with those of geodesic active contours. Experimental results demonstrate the superior performance of the proposed method.

• Proceedings of the Acoustical Society of Korea Conference
• /
• spring
• /
• pp.82-85
• /
• 1999

본 연구에서는 연속음성인식 시스템의 성능 향상을 위하여 음성의 채록환경 및 데이터량 등을 고려한 효과적인 언어모델 작성방법을 제안하고, 이를 항공편 예약시스템에 적용하여 성능 평가 실험을 실시한 결과 $91.6\%$의 인식률을 얻어 제안한 방법의 유효성을 확인하였다. 이를 위하여 소량의 200문장의 항공편 예약 텍스트 데이터를 이용하여 좀더 강건한 단어발생 확률을 가지도록 하기 위해 일반적으로 대어휘 연속음성인식에서 많이 이용되고 있는 단어 N-gram 언어모델을 도입하고 이를 다양한 발성환경을 고려하여 1,154문장으로 확장한 후 동일 문장'을 반복 학습하여 언어모델을 작성하였다. 인식에 있어서는 오인식과 문법적 오류를 최소화하기 위하여 forward - backward pass 방법의 stack decoding알고리즘을 이용하였다. 인식실험 결과, 평가용 3인의 200문장을 각 반복학습 회수에 따라 학습한 각 언어모델에 대해 평가한 결과, forward pass의 경우 평균 $84.1\%$, backward pass의 경우 평균 $91.6\%$의 문장 인식률을 얻었다. 또한, 반복학습 회수가 증가함에 따라 backward pass의 인시률의 변화는 없었으나, forward pass의 경우, 인식률이 반복회수에 따라 증가하다가 일정값에 수렴함을 알 수 있었고, 언어모델의 복잡도에서도 반복회수가 증가함에 따라 서서히 줄어들며 수렴함을 알 수 있었다. 이상의 결과로부터 소량의 텍스트 데이터를 이용한 제한된 태스크에서 언어모델을 작성할 때 반복학습 방법이 유효함을 확인할 수 있다.

• Convergence Security Journal
• /
• v.24 no.2
• /
• pp.201-207
• /
• 2024

Voice is a critical element of human communication, and the development of speech recognition models is one of the significant achievements in artificial intelligence, which has recently been applied in various aspects of human life. The application of speech recognition models in the military field is also inevitable. However, before artificial intelligence models can be applied in the military, it is necessary to research their vulnerabilities. In this study, we evaluates the military applicability of the multilingual speech recognition model "Whisper" by examining its vulnerabilities to battlefield noise, white noise, and adversarial attacks. In experiments involving battlefield noise, Whisper showed significant performance degradation with an average Character Error Rate (CER) of 72.4%, indicating difficulties in military applications. In experiments with white noise, Whisper was robust to low-intensity noise but showed performance degradation under high-intensity noise. Adversarial attack experiments revealed vulnerabilities at specific epsilon values. Therefore, the Whisper model requires improvements through fine-tuning, adversarial training, and other methods.

• Journal of Navigation and Port Research
• /
• v.47 no.4
• /
• pp.256-270
• /
• 2023

In this study, a sliding mode (SM) controller for dynamic positioning (DP) was specifically designed for a turret connection operation of a ship or an offshore structure in which an arbitrary point on the structure could be controlled as the motion center instead of the center of mass. The SM controller allows control of the arbitrary point and provides capability to manage uncertainties in the dynamics of ships and offshore structures, external forces caused by unknown changing marine environments, and transient performance of DP systems. The Jacobian matrix included in kinematic equations of the controlled object was modified to design the SM controller to control based on an arbitrary point of ships or offshore structures. To ensure robustness of the controller, the Lyapunov stability theory was applied in the design of the SM controller. In general, for robustness in DP control, gain scheduling based on a proportional-derivative (PD) control algorithm is employed. However, finding appropriate gains for gain scheduling complicates the application of DP systems. Therefore, in this study, the SM control algorithm was considered to mitigate the complexity of the DP controller for ships and offshore structures. To validate the proposed SM control algorithm, time-domain simulations were conducted and utilized to evaluate the performance of the control algorithm. The effectiveness of the proposed SM controller was assessed by comparing simulation results with results of a conventional PD control algorithm applied in DP control.

• Journal of the Korea Society for Simulation
• /
• v.27 no.1
• /
• pp.75-90
• /
• 2018

Generally, solutions that can increase patients' satisfaction are in conflict with the solutions that can increase the satisfaction and profitability of the medical personnels. In this paper, we compare appointment patterns using multiple performance measures that take into account both hospital's position and customers' position, and propose the best appointment pattern especially for an efficient scheduling in a plastic surgery clinic. A simulation model using ARENA is designed based on real hospital data. Based on this model, we compare 5 appointment patterns. To check the robustness of the appointment patterns, we have experimented on various appointment situations by considering peak, extreme peak seasons, and peak, off-peak days. We found that Triangle-like pattern turned out to be most efficient regardless of appointment situations.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2007.05a
• /
• pp.107-110
• /
• 2007

클라이언트-서버 구조의 스트리밍 시스템은 서버의 가용 능력에 따라 클라이언트의 개수가 제한되는 단점을 가진다. 본 논문에서는 인터넷 방송 시스템의 확장성과 안정성을 지원하기 위해 P2P에 기반한 모델을 제시한 후, 프로토타입 시스템을 설계 및 구현하였다. 본 논문에서 구현한 시스템 OmniCast264는 H.264 비디오 스트림을 제공하는 인코딩서버와 스트리밍서버, 피어노드, 프록시 서버로 구성된다. OmniCast264는 스트리밍 부하의 분산화, 실시간성, 에러 발생에 따른 강건함, 계층의 모듈화등의 개념을 가지고 있기 때문에 대용량의 인터넷 방송에 적합하다고 할 수 있다. 마지막으로 PC들을 병렬 및 직렬 구조의 유형으로 나누어 P2P를 구성한 후, OmniCast264의 성능을 평가하여 실시간 재생이 가능함을 검증하였다.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.34 no.1
• /
• pp.71-82
• /
• 2024

With the advancement of information and communication technology (ICT), the use of electronic document types such as PDF, MS Office, and HWP files has increased. Such trend has led the cyber attackers increasingly try to spread malicious documents through e-mails and messengers. To counter such attacks, AI-based methodologies have been actively employed in order to detect malicious document files. The main challenge in detecting malicious HWP(Hangul Word Processor) files is the lack of quality dataset due to its usage is limited in Korea, compared to PDF and MS-Office files that are highly being utilized worldwide. To address this limitation, data augmentation have been proposed to diversify training data by transforming existing dataset, but as the usefulness of the augmented data is not evaluated, augmented data could end up harming model's performance. In this paper, we propose an effective semi-supervised learning technique in detecting malicious HWP document files, which improves overall AI model performance via quantifying the utility of augmented data and filtering out useless training data.