• 한국음향학회지
• /
• 제18권5호
• /
• pp.44-51
• /
• 1999

본 논문에서는 음성과 자연언어의 통합처리를 위한 효과적인 병렬계산모델을 제안한다. 음소모델은 연속 Hidden Markov Model(HMM)에 기반을 둔 문맥종속형 음소를 사용하며, 언어모델은 지식베이스를 기반으로 한다. 또한 지식베이스를 구성하기 위해 계층구조의 semantic network과 병렬 marker-passing을 추론 메카니즘으로 쓰는 memory-based parsing 기술을 사용한다. 본 연구의 병렬 음성인식 알고리즘은 분산메모리 MIMD(Multiple Instruction Multiple Data) 구조의 다중 Transputer 시스템을 이용하여 구현되었다. 실험결과, 본 연구의 지식베이스 기반 음성인식 시스템의 인식률이 word network 기반 음성인식 시스템보다 높게 나타났으며 code-phoneme 통계정보를 활용하여 인식성능의 향상도 얻을 수 있었다. 또한, 성능향상도(speedup) 관련 실험들을 통하여 병렬 음성인식 시스템의 실시간 구현 가능성을 확인하였다.

• 대한산업공학회지
• /
• 제29권4호
• /
• pp.312-320
• /
• 2003

Recently there is a growing interest in mobile cellular network providing multimedia service. However, the link bandwidth of mobile cellular network is not sufficient enough to provide satisfactory services to users. To overcome this problem, an adaptive framework has been proposed. In this study, we propose a new method of estimating DPR(Degradation Period Ratio) in an adaptive multimedia network where the bandwidth of ongoing call can be dynamically adjusted during its lifetime. DPR is a QoS(Quality of Service) parameter which represents the ratio of allocated bandwidth below a pre-defined target to the whole service time of a call. We improve estimation method of DPR using DTMC(Discrete Time Markov Chain) model by calculate mean degradation period, degradation probability more precisely than in existing studies. Under Threshold CAC(Call Admission Control) algorithm, we present analytically how to guarantee QoS to users and illustrate the method by numerical examples. The proposed method is expected to be used as one of CAC schemes in guaranteeing predefined QoS level of DPR.

• 한국음향학회지
• /
• 제39권2호
• /
• pp.131-136
• /
• 2020

음성 인식기를 대기모드에서 동작 모드로 전환하기 위해 발화하는 짧은 단어를 기동어(Wake Up Word, WUW)라고 하며, 음성 인식기를 실제로 사용하는 사용자가 지정한 기동어를 사용자 정의 기동어라고 한다. 본 논문에서는 이러한 사용자 정의 기동어를 인식하기 위해 기존의 Gaussian Mixture Model-Hidden Markov Model(GMM-HMM) 기반의 시스템, Linear Discriminant Analysis(LDA)를 적용한 LDA-GMM-HMM 기반의 시스템과, LDA-GMM-HMM 모델에서 GMM을 Deep Neural Network(DNN)로 대체한 LDA-DNN-HMM 기반의 시스템을 제작하고 각 시스템의 사용자 정의 기동어 인식 성능 및 비기동어 거절 성능을 비교한다. 또한 기동어 인식기의 체감 성능을 향상시키고자 각 모델에 threshold를 적용하여 기동어 인식 실패율을 약 10 % 수준으로 감소 시킨 후에 비기동어(non-WUW)의 거절 실패율을 비교 평가한다. Threshold 적용시에 LDA-DNN-HMM 기반의 시스템의 경우 기동어 인식 실패율 9.84 % 수준에서 비기동어 거절 실패율이 0.0058 %의 인식 성능을 나타내어 LDA-GMM-HMM 시스템 보다 약 4.82배 향상된 비기동어 거절 성능을 나타낸다. 이러한 결과는 본 논문에서 제작한 LDA-DNN-HMM 모델이 사용자 정의 기동어 인식 시스템을 구축하는데 효과적임을 입증한다.

• 한국소음진동공학회논문집
• /
• 제23권9호
• /
• pp.814-822
• /
• 2013

For the fault diagnosis of a mechanical system, pattern recognition methods have being used frequently in recent research. Hidden Markov model(HMM) and artificial neural network(ANN) are typical examples of pattern recognition methods employed for the fault diagnosis of a mechanical system. In this paper, a hybrid method that combines HMM and ANN for the fault diagnosis of a mechanical system is introduced. A rotating blade which is used for a wind turbine is employed for the fault diagnosis. Using the HMM/ANN hybrid model along with the numerical model of the rotating blade, the location and depth of a crack as well as its presence are identified. Also the effect of signal to noise ratio, crack location and crack size on the success rate of the identification is investigated.

• 한국건축시공학회지
• /
• 제11권1호
• /
• pp.91-99
• /
• 2011

Bayesian network is a form of probabilistic graphical model. It incorporates human reasoning to deal with sparse data availability and to determine the probabilities of uncertain cases. In this research, bayesian network is adopted to model the problem of construction project cost. General information, time, cost, and material, the four main factors dominating the characteristic of construction costs, are incorporated into the model. This research presents verify a model that were conducted to illustrate the functionality and application of a decision support system for predicting the costs. The Markov Chain Monte Carlo (MCMC) method is applied to estimate parameter distributions. Furthermore, it is shown that not all the parameters are normally distributed. In addition, cost estimates based on the Gibbs output is performed. It can enhance the decision the decision-making process.

• 제어로봇시스템학회논문지
• /
• 제8권4호
• /
• pp.288-291
• /
• 2002

We characterize multipath fading channel dynamics at the packet level and analyze the corresponding data queueing performance in various environments. We identify the similarity between wire-line queueing analysis and wireless network per-formance analysis. The second order channel statistics, i.e. channel power spectrum, is fecund to play an important role in the modeling of multipath fading channels. However, it is identified that the first order statistics, i.e. channel CDF also has significant impact on queueing performance. We use a special Markov chain, so-called CMPP, throughout this paper.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2005년도 제36회 하계학술대회 논문집 A
• /
• pp.226-228
• /
• 2005

This paper deals with power system stabilization problem using a network control system in which the control is applied through a communication channel in feedback form. Analysis and synthesis issues are investigated by modeling the packet delivery characteristics of the network as a Bernoulli random variable, which is described by a two state Markov chain. This model assumption yields an overall system which is described by a discrete-time Markov jump linear system. These employ the norm to measure the performance of the system, and they compute the norm via a necessary and sufficient matrix inequality condition.

• 한국컴퓨터정보학회논문지
• /
• 제16권2호
• /
• pp.93-99
• /
• 2011

본 논문에서는 강화학습(RL : Reinforcement Learning) 환경 하에서 정책 기울기 값 기법을 사용하는 적응적인 QoS 라우팅 기법을 제안하였다. 이 기법은 기존의 강화학습 환경 하에 제공하는 기법에 비해 기대 보상값의 기울기 값을 정책에 반영함으로써 빠른 네트워크 환경을 학습함으로써 보다 우수한 라우팅 성공률을 제공할 수 있는 기법이다. 이를 검증하기 위해 기존의 기법들과 비교 검증함으로써 그 우수성을 확인하였다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제10권7호
• /
• pp.3171-3191
• /
• 2016

Due to growing demand on wireless data traffic, a large number of different types of base stations (BSs) have been installed. However, space-time dependent wireless data traffic densities can result in a significant number of idle BSs, which implies the waste of power resources. To deal with this problem, we propose an active state control algorithm based on semi-Markov decision process (SMDP) for a heterogeneous network. A MDP in discrete time domain is formulated from continuous domain with some approximation. Suboptimal on-line learning algorithm with a random policy is proposed to solve the problem. We explicitly include coverage constraint so that active cells can provide the same signal to noise ratio (SNR) coverage with a targeted outage rate. Simulation results verify that the proposed algorithm properly controls the active state depending on traffic densities without increasing the number of handovers excessively while providing average user perceived rate (UPR) in a more power efficient way than a conventional algorithm.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제7권4호
• /
• pp.285-294
• /
• 2007

In this paper, we investigate a novel online estimation algorithm for dynamic Bayesian network(DBN) parameters, given as conditional probabilities. We sequentially update the parameter adjustment rule based on observation data. We apply our algorithm to two well known representations of DBNs: to a first-order Markov Chain(MC) model and to a Hidden Markov Model(HMM). A sliding window allows efficient adaptive computation in real time. We also examine the stochastic convergence and stability of the learning algorithm.