• Proceedings of the Acoustical Society of Korea Conference
• /
• 1998.06c
• /
• pp.71-74
• /
• 1998

HMM (hidden Markov model)을 이용한 음성인식은 현재 가장 널리 쓰여지고 있는 방법으로, 이 중 CDHMM (continuous observation density HMM)은 상태에서 관측심볼확률을 연속확률밀도를 사용하여 표현한다. 본 논문에서는 가우스 혼합밀도함수를 사용하는 CDHMM의 상태천이확률과, 관측심볼확률을 표현하기 위한 인자인 평균벡터, 공분산 행렬, 가지하중값을 유전자 알고리듬을 사용하여 최적화하는 방법을 제안하였다. 유전자 알고리듬은 매개변수 최적화문제에 대하여 자연의 진화원리를 모방한 알고리듬으로, 염색체 형태로 표현된 개체군 (population) 중에서 환경에 대한 적합도 (fitness)가 높은 개체가 높은 확률로 살아남아 재생 (reproduction)하게 되며, 교배 (crossover)와 돌연변이 (mutation) 연산 후에 다음 세대 개체군을 형성하게 되고, 이러한 과정을 반복하면서 최적의 개체를 구하게 된다. 본 논문에서는 상태천이확률, 평균벡터, 공분산행렬, 가지하중값을 부동소수점수 (floating point number)의 유전자형으로 표현하여 유전자 알고리듬을 수행하였다. 유전자 알고리듬은 복잡한 탐색공간에서 최적의 해를 찾는데 효과적으로 적용되었다.

• Proceedings of the Korea Multimedia Society Conference
• /
• 2000.04a
• /
• pp.49-52
• /
• 2000

멀티미디어 데이터 저장에 있어서 중요한 요소 중의 하나는 얼마나 많은 양의 정보를 주어진 공간에 기록할 수 있느냐는 것이다. 현재 많이 사용되고 있는 광저장장치(optical channel) 기록을 위하여 Squared M-ary RLL의(d, k) 제한조건을 이용하여 상태천이행렬을 일반화하고 채널용량을 구하였다.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.4 no.1
• /
• pp.3-12
• /
• 1994

본 연구는, 복잡한 비선형시스템의 표현에 뛰어난 능력을 갖고 있는 TSK형 퍼지모델로 부터 전체시스템의 안정성을 보장할 수 있는 퍼지상태제어기의 설계방법을 제안한다. 그 퍼지상태제어기는 TSK형 퍼지모델과 같은 형태의 퍼지규칙들로 구성되며, 추론 방법 및 상태제어 파라미터 행렬을 구하는 방벙은 전체시스템의 상태천이행렬이 원하는 안정한 것이 될 수 있도록 정해진다. 또한 본 논문에서는, 현재 대부분이 숙련가에 의해 수동으로 조작되고 있는 컨테이너 크래인의 새로은 제어 방법을 제시하고 제안한 퍼지상태제어기를 적용한다.컨테이너 크래인의 모형을 만들어, 제어숙련가의 수동조작으로 결정되는 트롤리와 승강기의 규범속도를 퍼지모델로 표현하고, 트롤리와 승강기가 그 규범 속도에 따르도록 제안한 퍼지상태제어기로 제어한다. 제안된 방법을 실험한 결과 모형크레인의 궤적이 숙련가에 의해 만들어진 궤적과 매우 유사하게 됨을 알 수 있었다.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.12 no.4
• /
• pp.575-584
• /
• 2017

It is suggested the circuit network to analyze the time response of control system. And it is analyzed the interrelation for classical control and modern control by the transfer function and the state equation. Without complicated integration of state transition equation, it is suggested to possible time response by combining the state transition matrix and the transfer function. A source program is coded to display the time response according to the unit-step and the sinusoidal input. Transient response is analyzed in the unit-step input and phase difference between current and voltage is analyzed in sinusoidal input. As writing the suggested contents in transient response or state-space analysis, it is improved the understanding for control engineering and ability for system design.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.46 no.4
• /
• pp.39-48
• /
• 2009

Asynchronous sequential machines, or clockless logic circuits, have several advantages over synchronous machines such as fast operation speed, low power consumption, etc. In this paper, we propose a novel robust controller for input/output asynchronous sequential machines with uncertain state transitions. Due to model uncertainties or inner failures, the state transition function of the considered asynchronous machine is not completely known. In this study, we present a formulation to model this kind of asynchronous machines ana using generalized reachability matrices, we address the condition for the existence of an appropriate controller such that the closed-loop behavior matches that of a prescribed model. Based on the previous research results, we sketch design procedure of the proposed controller and analyze the stable-state operation of the closed-loop system.

• The Journal of the Acoustical Society of Korea
• /
• v.13 no.4
• /
• pp.40-49
• /
• 1994

In this paper, we propose methods for improving the performance of word recognition system. The ray stratey of the first method is to apply the inertia to the feature vector sequences of speech signal to stabilize the transitions between VQ cdoes. The second method is generating the new observation probabilities using the transition matrix of VQ codes as weights at the observation probability of the output symbol, so as to take into account the time relation between neighboring frames in DHMM. By applying the inertia to the feature vector sequences, we can reduce the overlapping of probability distribution of the response paths for each word and stabilize state transitions in the HMM. By using the transition matrix of VQ codes as weights in conventional DHMM. we can divide the probability distribution of feature vectors more and more, and restrict the feature distribution to a suitable region so that the performance of recognition system can improve. To evaluate the performance of the proposed methods, we carried out experiments for 50 DDD area names. As a result, the proposed methods improved the recognition rate by $4.2\%$ in the speaker-dependent test and $12.45\%$ in the speaker-independent test, respectively, compared with the conventional DHMM.

• Journal of Korean Society of Forest Science
• /
• v.81 no.4
• /
• pp.325-336
• /
• 1992

The successional status and potential natural vegetation were examined in a mixed broadleaved-Abies forest located at Mt, Odae National Park. The examination was based on the subsequent process of generation replacement by understory saplings for the dominant canopy trees within 88 $5m{\times}50m$ belt transects. The transition matrix model, which was modified from mathematical theory of Markov chain, was employed to analyze the successional status of the study forest. The model suggests that the study forest is still seral, and it is considered to be more than 700 years away from the steady state or climax in terms of species composition. The simulations predict a remarkable decrease in the proportion of species composition of the present dominant Quercus mongolica and Tilia amurensi.s from current 28% and 13% to less than 3% and 5%, respectively, at the steady state. On the contrary, the proportions of Abies holophylla, Acer pseudo-sieboldianum, Carpinus cordata, and Pinus koraiensis will increase at the steady state. The ratio of mixture between deciduous and coniferous trees will be gradually changed from current 6.5 : 3.5 to 5.0 : 5.0 at equilibrium. Discussion for the problems of Quercus mongolica in successional status noted that the species behaved as a mid-successional dominant, not a climax species in the study forest. The hypothesis and sensitivity of the transition matrix model were also discussed to evaluate the applicability to the real situation. The overall results indicated that the present dynamics of the forest must reflect the seral state due to previous disturbance mainly by hyman related interference.

• Journal of Korean Society of Forest Science
• /
• v.94 no.6
• /
• pp.387-396
• /
• 2005

Having employed the transitional probability model based on Markov chain, the study was carried out to examine successional trends for community types in the natural deciduous forest of Mt. Jumbong. The species composition of oncoming generation in overstory was estimated from that of mid-story, and the species composition in mid-story was based upon that of understory. Successional trend for each community was predicted from the reorganized probability matrix of tree replacement by the square of climax index, which was evaluated by the factors of light absorption, reproduction, and wood quality. As the result of analysis, following table shows the oncoming generation of steady state and dominant species in overstory and mid-story by community types. Even though Acer pseudo-sieboldianum and Carpinus cordata could hardly reach the canopy layer due to the intrinsic growth form, these species were predicted to maintain high compositional ratio so as to play an important ecological role in the study forest ecosystem.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.19 no.5
• /
• pp.805-814
• /
• 1994

In this paper, we introduce a new method which is available for analyzing the throughput of the packet radio network by using the auxiliary Markov transient matrix with a failure state and a success state. And we consider the effect of symbol error for the network state(X, R) consisted of the number of transmitting PRU X and receiving PRU R. We examine the packet radio network of a continuous time Markov chain model, and the direct sequence binary phase shift keying CDMA radio channel with hard decision Viterbi decoding and bit-by-bit changing spreading code. For the unslotted distributed multi-hop packet radio network, we assume that the packet error due to a symbol error of radio channel has Poisson process, and the time period of an error occurrence is exponentially distributed. Through the throughputs which are found as a function of radio channel parameters, such as the received signal to noise ratio and chips of spreading code per symbol, and of network parameters, such as the number of PRU and offered traffic rate, it is shown that this composite analysis enables us to combine the Markovian packet radio network model with a coded DS/BPSK CDMA radio channel.

• Journal of Astronomy and Space Sciences
• /
• v.20 no.1
• /
• pp.21-28
• /
• 2003

In case that we independently obtain orbital informations about the low earth satellites of foreign countries using radar systems, we develop the orbit determination algorithm for this purpose using a SGP4 model with an analytical orbit model and the extended Kalman filter with a real-time processing method. When the state vector is Keplerian orbital elements, singularity problems happen to compute partial derivative with respect to inclination and eccentricity orbit elements. To cope with this problem, we set state vector osculating to mean equinox and true equator cartesian elements with coordinate transformation. The state transition matrix and the covariance matrix are numerically computed using a SGP4 model. Observational measurements are the type of azimuth, elevation and range, filter process to each measurement in a lump. After analyzing performance of the developed orbit determination algorithm using TOPEX/POSEIDON POE(precision 0.bit Ephemeris), its position error has about 1 km. To be similar to performance of NORAD system that has up to 3km position accuracy during 7 days need to radar system performance that have accuracy within 0.1 degree for azimuth and elevation and 50m for range.