• Journal of Aerospace System Engineering
• /
• v.11 no.4
• /
• pp.22-29
• /
• 2017

A study is performed for the effective detection method of a fault which is occurred during operation in a small turbojet engine with non-linear characteristics used by unmanned air vehicle. For this study the non-linear dynamic model of the engine is derived from transient thermodynamic cycle analysis. Also for inducing real operation conditions the controller is developed associated with unscented Kalman filter to estimate noises. Sequential probability ratio test is introduced as a real time method to detect a fault which is manipulated for simulation as a malfunction of rotational speed sensor contaminated by large amount of noise. The method applied to the fault detection during operation verifies its effectiveness and high feasibility by showing good and definite decision performances of the fault.

• Proceedings of the Acoustical Society of Korea Conference
• /
• spring
• /
• pp.75-78
• /
• 2000

본 논문은 자연스러운 합성음 생성을 위한 끊어 읽기 강도 예측에 관한 것으로, 문장에 대한 품사열이 주어졌을 때 Posteriori 확률을 최대화하는 끊어 읽기 강도를 비터비 디코딩으로 예측한다. 훈련용 데이터는 여성화자 1인이 발성한 2,100 문장이며, 음성 데이터로부터 휴지길이(pause)에 따라 끊어 읽기 강도를 2단계로 할당하고, 텍스트에서는 30개의 품사 태그 심볼을 이용하여 형태소분석 및 태깅을 수행하였다. 관측확률은 3개 연속하는 품사열이 발생할 확률로 하고 끊어 읽기 강도 천이확률은 bigram으로 했을 때, cross validation 방법으로 성능 평가를 수행하였다 평가결과, 훈련데이타에 대해서는 $89.7\%$, 테스트 데이터에 대해서는 $84.9\%$의 예측정확률을 보였다.

• Proceedings of the Safety Management and Science Conference
• /
• 2010.11a
• /
• pp.331-335
• /
• 2010

본 논문에서는 위치기반서비스의 핵심기능을 담당하는 측위기술 중 주변의 다양한 잡음환경에서도 다른 측위기술에 비해 비교적 측위오차를 줄일 수 있는 Fingerprint 기법을 적용하여 객체의 위치를 결정하는 방법을 제시하였다. 특히, Fingerprint 기법을 활용함에 있어서 결정적 측위방법이 아닌 확률적 측위방법을 사용하였고, 기존의 연구들에서 제시된 확률적 측위방법에서 인접 RP(Reference Point)들과의 영향성까지 고려한 측위방법을 제시한 것이 본 연구의 특징이다. 비주얼베이직 6.0으로 측위 프로그램을 구현하였고, 실내 외 연속측위의 선행연구로서 실내 측위 테스트를 통해 기존의 결정적 방법, 확률적 방법, 그리고 본 연구에서 제안하는 방법의 측위 정확성을 평가하였다. 테스트 결과, 결정적 측위 방법 및 기존의 확률적 측위방법과 비교하여 각각 평균 15%와 7% 측위 오차가 감소하였다.

• 한국HCI학회:학술대회논문집
• /
• 2008.02a
• /
• pp.245-250
• /
• 2008

최근 동작 및 행동 인식에 대한 연구가 활발하다. 특히, 센서가 소형화되고 저렴해지면서 그 활용을 위한 관심이 증가하고 있다. 기존의 많은 행동 인식 연구에서 사용되어 온 정적 분류 기술 기반 동작 인식 방법은 연속적인 데이터 분류 기술에 비해 유연성 및 활용성이 부족할 수 있다. 본 논문에서는 연속적인 데이터의 패턴 분류 및 인식에 효과적인 확률적 추론 기법인 은닉 마르코프 모델(Hidden Markov Model)과 사전 지식 없이도 자동 학습이 가능하며 의미 깊은 궤적 패턴을 클러스터링하고 효과적인 양자화가 가능한 자기구성지도(Self Organizing Map)를 이용한 동작 인식 기술을 소개한다. 또한, 그 유용성을 입증하기 위해 실제 가속도 센서를 이용하여 다양한 동작에 대한 데이터를 수집하고 분류 성능을 분석 및 평가한다. 실험에서는 실제 가속도 센서를 통해 수집된 숫자를 그리는 동작의 성능 평가 결과를 보이고, 행동 인식기 별 성능과 전체 인식기별 성능을 비교한다.

• Journal of the Korean Society of Hazard Mitigation
• /
• v.7 no.5
• /
• pp.79-88
• /
• 2007

This study evaluated the statistical stationarity of rainfall quantiles as well as the rainfall itself. The conventional methodologies like the Cox-Stuart test for trend and Dickey-Fuller test for a unit root used for testing the stationarity of a time series were applied and evaluated their application to the rainfall quantiles. As results, first, no obvious increasing or decreasing trend was found for the rainfall in Seoul, which was also found to be a stationary time series based on the Dickey-Fuller test. However, the Cox-Stuart test for the rainfall quantiles show some trends but not in consistent ways of increasing or decreasing. Also, the Dickey-Fuller test for a unit root shows that the rainfall quantiles are non-stationary. This result is mainly due to the difference between the rainfall data and rainfall quantiles. That is, the rainfall is a random variable without any trend or non-stationarity. On the other hand, the rainfall quantiles are estimated by considering all the data to result in high correlation between their consecutive estimates. That is, as the rainfall quantiles are estimated by adding a stationary rainfall data continuously, it becomes possible for their consecutive estimates to become highly correlated. Thus, it is natural for the rainfall quantiles to be decided non-stationary if considering the methodology used in this study.

• Journal of the Earthquake Engineering Society of Korea
• /
• v.14 no.3
• /
• pp.59-68
• /
• 2010

Seismic performances of existing structures should be assessed with more accuracy for cost-effective retrofits. Existing bridges are assessed by the current guidelines in which a simple method has been adapted considering the technical level of engineers of the historical time of construction. Recently many probabilistic approaches have been performed to reflect the uncertainties of seismic input motions. Structures are modeled frequently with the neglection of soil foundations or modeled occasionally with elastic soil spring elements to consider the effect of the soil on the structural response. However, soil also shows nonlinearity under seismic events, so this characteristic should be reflected in order to obtain a more accurate assessment. In this study, a 6-span continuous bridge has been analyzed under various seismic events, in which the soil was represented by equivalent linear spring elements having different properties according to the intensities of the input motions experienced. The seismic vulnerabilities with respect to the failure of piers and the dropping of the super-structure were evaluated on the basis of the analysis results.

• Composites Research
• /
• v.21 no.6
• /
• pp.23-30
• /
• 2008

In this paper, the effect of the natural aging on the strength distribution and structural service life of a Filament Wound (FW) composite pressure vessel was studied. The fiber failure strain, which is varied significantly, was considered as the design random variable and the strength analysis was carried out by probabilistic numerical approach. The progressive failure analysis technique and the First Order Reliability Method (FORM) were embedded in this numerical model. As the calculation results, the probability of failure was obtained for each aging time steps and it is found that the strength degradation in FW composite pressure vessel, due to the natural aging, appears within 10 year-aging-time. As an example of the life prediction under natural aging using arbitrary laminated model, the service lifetime of 13 years was predicted based on the probability of failure of 2.5% and the design pressure of 3,250 psi.

• Journal of the Korea institute for structural maintenance and inspection
• /
• v.23 no.2
• /
• pp.10-19
• /
• 2019

Management plan with repairing is essential for RC structures exposed to chloride attack since durability problems occur with extended service life. Conventionally deterministic method is adopted for evaluation of service life and repair cost, however more reasonable repair cost can be obtained through continuous repair cost from probabilistic maintenance technique. Unlike the previous researches considering only normal distribution of life time, PLTFs (Probabilistic Life Time Function) which can be capable of handling log and normal distributions are attempted for initial and repair service life, and repair cost is evaluated for OPC and GGBFS concrete. PLTF with log distributions in initial service life is more effective to save repair cost since it is more dominant after average than normal distribution. Repair cost in GGBFS concrete decreases to 30% of OPC concrete due to longer initial service life and lower repairing event. The proposed PLTF from the work can handle not only normal distributions but also log distributions for initial and repair service life, so that it can provide more reasonable repair cost evaluation.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.7B
• /
• pp.660-666
• /
• 2004

Overall TCP performance represented by end-to-end throughput is largely dependent upon its loss recovery performance. In particular non-congestion packet losses caused by transmission errors degrade TCP performance seriously. Using Markov process, we analyze TCP loss recovery performance for correlated packet losses caused by multipath fading. The results show that loss recovery performance can be severely affected by burstiness in packet losses, even if overall packet loss ratio is very low.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.30 no.4B
• /
• pp.389-397
• /
• 2010

This study introduced a Bayesian based frequency analysis in which the statistical trend analysis for hydrologic extreme series is incorporated. The proposed model employed Gumbel extreme distribution to characterize extreme events and a fully coupled bayesian frequency model was finally utilized to estimate design rainfalls in Seoul. Posterior distributions of the model parameters in both Gumbel distribution and trend analysis were updated through Markov Chain Monte Carlo Simulation mainly utilizing Gibbs sampler. This study proposed a way to make use of nonstationary frequency model for dynamic risk analysis, and showed an increase of hydrologic risk with time varying probability density functions. The proposed study showed advantage in assessing statistical significance of parameters associated with trend analysis through statistical inference utilizing derived posterior distributions.