• 제어로봇시스템학회논문지
• /
• 제19권4호
• /
• pp.306-314
• /
• 2013

Automatic tracking in cluttered environment requires the initiation and maintenance of tracks, and track existence probability of true track is kept by Markov Chain Two model of target existence propagation. Unlike Markov Chain One model for target existence propagation, Markov Chain Two model is made up three hypotheses about target existence event which are that the target exist and is detectable, the target exists and is non-detectable through occlusion, and the target does not exist and is non-detectable according to non-existing target. In this paper we present multi-scan single target tracking algorithm based on the target existence, which call the Integrated Track Splitting algorithm with Markov Chain Two model in imaging sensor.

• 대한음성학회:학술대회논문집
• /
• 대한음성학회 2005년도 춘계 학술대회 발표논문집
• /
• pp.87-90
• /
• 2005

Today's state-of the-art speech recognition systems typically use continuous distribution hidden Markov models with the mixtures of Gaussian distributions. To obtain higher recognition accuracy, the hidden Markov models typically require huge number of Gaussian distributions. Such speech recognition systems have problems that they require too much memory to run, and are too slow for large applications. Many approaches are proposed for the design of compact acoustic models. One of those models is subspace distribution clustering hidden Markov model. Subspace distribution clustering hidden Markov model can represent original full-space distributions as some combinations of a small number of subspace distribution codebooks. Therefore, how to make the codebook is an important issue in this approach. In this paper, we report some experimental results on various quantization methods to make more accurate models.

• Journal of Biosystems Engineering
• /
• 제29권5호
• /
• pp.441-450
• /
• 2004

This study was performed to estimate mid and long term demands of a tractor, a rice transplanter, a combine and a grain dryer by using logistic curve function and Markov chain model. Field survey was done to decide some parameters far logistic curve function and Markov chain model. Ceiling values of tractor and combine fer logistic curve function analysis were 209,280 and 85,607 respectively. Based on logistic curve function analysis, total number of tractors increased slightly during the period analysed. New demand for combine was found to be zero. Markov chain analysis was carried out with 2 scenarios. With the scenario 1(rice price $10\%$ down and current supporting policy by government), new demand for tractor was decreased gradually up to 700 unit in the year 2012. For combine, new demand was zero. Regardless of scenarios, the replacement demand was increased slightly after 2003. After then, the replacement demand is decreased after the certain time. Two analysis of logistic owe function and Markov chain model showed the similar trend in increase and decrease for total number of tractors and combines. However, the difference in numbers of tractors and combines between the results from 2 analysis got bigger as the time passed.

• 대한기계학회논문집A
• /
• 제20권4호
• /
• pp.1241-1250
• /
• 1996

The characteristics of fatigue cumulative damage and fatigue life of 8-harness satin woven CFRP composites with a circular hole under constant amplitude and 2-level block loading are estimated by Stochastic Makov chain model. It is found in this study that the fatigue damage accumulation behavior is very random and the fatigue damage is accumulated as two regions under constant amplitude fatigue loading. In constant amplitude fatigue loading the predicted mean number of cycles to a specified damage state by Markov chain model shows a good agreement with the test result. The predicted distribution of the fatigue cumulative damage by Markov chain model is similar to the test result. The fatigue life predictions under 2-level block loading by Markov chain model revised are good fitted to the test result more than by 2-parameter Weibull distribution function using percent failure rule.

• 전자공학회논문지B
• /
• 제32B권12호
• /
• pp.1663-1669
• /
• 1995

This paper is a study on the recognition of Korean proverb using neural network and Markov model. The neural network uses, at the stage of training neurons, features such as the rate of zero crossing, short-term energy and PLP-Cepstrum, covering a time of 300ms long. Markov models were generated by the recognized phoneme strings. The recognition of words and proverbs using Markov models have been carried out. Experimental results show that phoneme and word recognition rates are 81. 2%, 94.0% respectively for Korean proverb recognition experiments.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2006년도 추계학술대회 논문집 전력기술부문
• /
• pp.357-359
• /
• 2006

Traditional time based preventive maintenance is used to constant maintenance interval for equipment life. In order to consider economic aspect for time based preventive maintenance, preventive maintenance is scheduled by RCM(Reliability-Centered Maintenance) evaluation. So, Markov state model is utilized considering stochastic state in RCM. In this paper, a Markov state model which can be used for scheduling and optimization of maintenance is presented. The deterioration process of system condition is modeled by a Markov model. In case study, simulation results about RCM are used to the real historical data of combustion turbine generating units in Korean power systems.

• 한국정보통신학회논문지
• /
• 제8권6호
• /
• pp.1175-1181
• /
• 2004

본 논문에서는 비디오 검색을 위한 비디오 사진 분류 시스템을 제안하였다. 제안된 시스템은 3개의 모듈인 특징 추출, 은닉 마르코브 모델 생성, 그리고 비디오 사진 분류로 구성되어 있다. 같은 등급에 속한 비디오 화면들이 반드시 유사하지 않으므로 견실한 Hidden Markov Model을 구성하기 위해서 는 충분한 학습이 필요하였다. 제안된 시스템은 텔레비전 야구 중계 방송의 비디오 화면을 15가지 등급으로 분류하여 분석 및 하는 실험을 한 결과 평균 84.72%의 인식률을 얻을 수 있었다.

• International journal of advanced smart convergence
• /
• 제7권1호
• /
• pp.15-23
• /
• 2018

Spectrum allocation is a key operation in cognitive radio networks (CRNs), where secondary users (SUs) are usually selfish - to achieve itself utility maximization. In view of this context, much prior lit literature proposed spectrum allocation base on non-cooperative game models. However, the most of them proposed non-cooperative game models based on complete information of CRNs. In practical, primary users (PUs) in a dynamic wireless environment with noise uncertainty, shadowing, and fading is difficult to attain a complete information about them. In this paper, we propose a non-cooperative game joint hidden markov model scheme for spectrum allocation in CRNs. Firstly, we propose a new hidden markov model for SUs to predict the sensing results of competitors. Then, we introduce the proposed hidden markov model into the non-cooperative game. That is, it predicts the sensing results of competitors before the non-cooperative game. The simulation results show that the proposed scheme improves the energy efficiency of networks and utilization of SUs.

• 한국경영과학회지
• /
• 제14권1호
• /
• pp.1-15
• /
• 1989

Multistate survival data with censoring often arise in biomedical experiments. In particular, a four-state space is used for cancer clinical trials. In a four-state space, each patient may either respond to a given treatment and then relapse or may progress without responding. In this four-state space, a model which combines the Markov and semi-Markov models is proposed. In this combined model, the generalized maximum likelihood estimators of the Markov and semi-Markov hazard functions are derived. These estimators are illustrated for the data collected in a study of treatments for advanced breast cancer.

• 한국산학기술학회논문지
• /
• 제15권11호
• /
• pp.6565-6575
• /
• 2014

본 논문에서 Hamilton의 마코프-스위칭 모형을 연립방정식으로 확장한 FIML 마코프-스위칭 모형을 제시해 보았다. 본 논문의 FIML 마코프-스위칭 모형을 LIML 마코프-스위칭 모형 등과 비교하면 LIML 마코프-스위칭 모형은 FIML 마코프-스위칭 모형의 특별한 경우이며 FIML 마코프-스위칭 모형은 연립방정식으로 확장된 일반화된 모형 형태를 띄게 된다. 본 논문의 FIML 마코프-스위칭 모형을 Campbell and Mankiw 소비함수에 적용해 본 결과, 2008년 부동산 거품 붕괴와 같은 경제충격 시기의 한계소비성향은 매우 민감도가 높아진다는 것을 알 수 있다.