• Proceedings of the Korea Information Processing Society Conference
• /
• 2011.11a
• /
• pp.269-272
• /
• 2011

일반적으로 지능형 에이전트에게 요구되는 가장 기초적인 상황 인식 기능 중의 하나가 불확실한 센서 데이터에 의존하여 자신의 현재 위치가 어디인지를 파악하는 일이다. 본 논문에서는 대표적인 확률기반의 측위 기법인 파티클 필터를 실제 로봇 측위에 적용한 실험을 수행하고, 이를 통해 측위 성능을 개선시킬 수 있는 방법들을 찾아본다. 특히 로봇 동작의 오차를 고려하지 않은 비-잡음 상태 전이 모델과 로봇 동작의 오차를 고려한 잡음 모델간의 비교 실험을 통해, 불확실성이 높은 실제 로봇 동작에 보다 근사한 상태 전이 모델이 파티클 필터 측위의 성능 개선에 도움이 될 수 있는지 분석해본다.

• Korean Journal of Social Welfare
• /
• v.69 no.3
• /
• pp.193-212
• /
• 2017

This study is purposed to explain the characteristics of injured workers' labor market situation and to analyze the factors influencing labor market transition of those workers. Using the Worker's Compensation Insurance Panel Data ver.1~3 which was surveyed by the Korean Workers' Compensation & Welfare Service in 2013-2015, this study analyzed 1,668 injured worker cases. The study shows that workers who have experience job retention at least once are 36.8% of all, 51.5% of them have experienced re-employment, and 36.9% have done unemployment. One result of the longitudinal analysis is that socio-demographic factors including gender, age, education years, convalescence period, ability on job performance, company size, term of service, temporary employment, daily-workers status before job accident and job training were associated with return to pre-injury job. The other result is that statistically significant factors affecting the probability to be the unemployed are gender, age, levels of disability, convalescence, ability on job performance, term of service before job accident, job rehabilitation service utilization. These findings indicate that we need to develop efficient intervention programs for supporting return-to-work and labor market transition of injured workers.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.17 no.1
• /
• pp.42-54
• /
• 2018

This study is to establish the traffic volume-based warrants of left-turn lanes in unsignalized intersections based on a risk probability methodology. This study applied a risk probability of a potential rear-end collision between a left-turn vehicle and the immediately following through vehicle. With the shifted negative exponential model and the compound probability theorem, the risk probability can be expressed as the function of directional volumes, opposing volumes and the percentage of left-turns for a two-lane and four-land highway, respectively. The warrants of installing left-turn lanes on unsignalized intersections were developed with the risk probability. The warrants define the total approaching and opposing volumes to encourage a left-turn lane as a function of operating speed, percentage of left-turn, and number of lanes.

• Journal of Korea Water Resources Association
• /
• v.35 no.5
• /
• pp.583-595
• /
• 2002

The purposes of this study are to perform the management and monitoring of droughts for Mokpo area via the monthly Palmer index(PDSI), the data is obtained from the Mokpo meteorological station, and the used data are in the period of 1906 to 1999. Monthly Palmer index is classified into 7 stochastic classes and its dynamic change of monthly transition probability estimated by Markov chain is investigated. We also estimate the steady state probability of the classified PDSI. The 4th class shows the highest frequency of 49.6% out of 7 classes and the 7th class which is the most extreme drought show that a stochastic transition probability is more or less larger than an empirical one. Also, we found that the monthly steady state probability could be used for the forecasting of changing pattern of drought magnitude for the study area.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 1996.04a
• /
• pp.491-494
• /
• 1996

주문이 매 시점마다 있는 것이 아니라 간헐적인, 즉 어느 시점에는 주문이 있고(ON) 다른시점에는 주문이 없는(OFF) 패턴에서 미래의 주문량에 대한 예측을 고려한다. 다음 시점의 예측량은 우선 주문이 있을 것인가에 대한 판단과 주문이 있다면 어느정도가 예상되는가 하는 문제의 두 가지 측면을 모두 고려해야 한다. 기존의 예측모델은 주문량 자체에 대한 고려가 일반적이며 주문시기에 대한 고려는 전무한 상태이기 때문에 이와 같은 주문패턴을 반영시키는데는 어려움이 따른다고 볼 수 있다. 본 논문에서는 이러한 주문패턴을 마코브 체인으로 모델링하고, 이러한 형태의 상태전이확률(state transition probaility) 추정식이 각각 독립적인 오목함수 (concave function)로 구성되어 있음을 보인다. 또한 확률적으로 표현되는 미래의 주문상태들에 대한 패턴을 확정시키는 알고리듬과 주문량 추정에 있어서 과거의 주문패턴을 반영시키는 모델을 제시한다.

• Annual Conference on Human and Language Technology
• /
• 1993.10a
• /
• pp.149-157
• /
• 1993

음성인식을 위해 다양한 방법들이 제안되어 있다. 본 연구에서는 음소단위 각각의 벡터 양자화된 코드북의 색인을 학습하는 HMM을 이용하여 한국어 숫자음을 대상으로 인식 실험을 수행하였다. 실험결과, 기존의 단어단위 HMM과 음소단위로 이루어진 유한상태기계(FSM)구조의 인식기에 비해 높은 인식율을 보였다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2000.10b
• /
• pp.371-373
• /
• 2000

본 논문에서는 분절 특징을 모수적 궤적 모델을 이용하여 표현하고, 이 특징을 분절 HMM(segmental HMM)의 입력으로 하는 음성 신호의 모델링 방식을 제안한다. 분절 특징은 음성의 경향을 나타내는 궤적으로 표현되고, 그 궤적은 연속되는 프레임 상에서 전이 정보를 포함하도록 디자인 행렬과 다항식의 회귀 함수를 이용하여 구해진다. 이 궤적을 분절 HMM에 적용하기 위하여, 외적 분절 변이와 내적 분절 변이에 대한 확률 분포 표현을 개선하였다. 제안된 방법의 효과를 살펴보기 위하여 TIMIT 데이터 베이스를 이용하여 실험한 결과, 제안된 분절 특징은 음성 신호의 인접한 프레임간의 상관관계를 표현하는 동적 특징과 같은 효과를 보였으며, 1차 미분계수를 포함하여 분절 특징을 구한 경우에는 기존의 특징 표현보다 좋은 성능을 보였다.

• Journal of Korea Water Resources Association
• /
• v.33 no.3
• /
• pp.307-314
• /
• 2000

This study is to investigate the recurrence characteristics of wet and dry years using over 200 year records of annual rainfall depth including Chosun Age in Korea. As well as analyzing the correlation structure of the raw data, recurrence trends of wet and dry year has been investigated based on several truncation levels (mean, $mean{\pm}0.25stdv.,\;mean{\pm}O.5stdv.,\;mean{\pm}O.75stdv.,\;mean{\pm}stdv.$). Also the transition probability among wet, dry and normal years has been derived for the same truncation levels. and finally the average return periods based on the steady-state probabilities were obtained. This analysis has been applied to not only the entire data but also partial data set of before- and after-the long dry period around 1900 in order to compare and detect the possible difference between the Chukwooki (an old raingauge invented in Chosun age) and the modem flip-bucket style. As a result, Similar pattern of dry and wet year recurrence has been found, but the return period of extremely dry years after the dry period shown longer than that before the dry period. Assuming that the dry and wet years can be defined as $mean{\pm}$ standard deviations, respectively, the return period of the wet years is shown to be about 5~6 years and that of the dry years about 6~7 years.

• Journal of the Korea Convergence Society
• /
• v.12 no.4
• /
• pp.229-236
• /
• 2021

There has been a lot of research on the real estate market, but a lack of research on the supply and demand of housing supply in each region, reflecting the changes in population growth and supply. It is calculated as the transition probability of the Markov chain model by reflecting the data on the number of houses per 1,000 people in the past 35 years and the forecast data for population change by region, in terms of supply (housing) to demand (population) for factors on the real estate market. According to the calculation results of the real estate market by region, the housing supply to the metropolitan area such as Gyeong-gi, Incheon, and Seoul is expected to be insufficient for a considerable period of time, considering the population changes by region. To stabilize the real estate market, it was confirmed that it was necessary to actively apply the differentiation of housing supply by region. It is meaningful in terms of verifying long term trends in the real estate market by region that reflect the prediction of population change, and it is expected that the methods used in this study will be practical through the analysis results using the historical data.

• Journal of KIISE:Computing Practices and Letters
• /
• v.13 no.7
• /
• pp.488-498
• /
• 2007

This paper presents a novel approach for representing the spatio-temporal topology of the camera network with overlapping and non-overlapping fields of view (FOVs) in Ubiquitous Smart Space (USS). The topology is determined by tracking moving objects and establishing object correspondence across multiple cameras. To track people successfully in multiple camera views, we used the Merge-Split (MS) approach for object occlusion in a single camera and the grid-based approach for extracting the accurate object feature. In addition, we considered the appearance of people and the transition time between entry and exit zones for tracking objects across blind regions of multiple cameras with non-overlapping FOVs. The main contribution of this paper is to estimate transition times between various entry and exit zones, and to graphically represent the camera topology as an undirected weighted graph using the transition probabilities.