• Journal of Digital Convergence
• /
• v.13 no.1
• /
• pp.243-248
• /
• 2015

Recognition model is not defined when you configure a model, Been added to the model after model building awareness, Model a model of the clustering due to lack of recognition models are generated by modeling is causes the degradation of the recognition rate. In order to improve decision tree state tying modeling using parameter estimation of Bayesian method. The parameter estimation method is proposed Bayesian method to navigate through the model from the results of the decision tree based on the tying state according to the maximum probability method to determine the recognition model. According to our experiments on the simulation data generated by adding noise to clean speech, the proposed clustering method error rate reduction of 1.29% compared with baseline model, which is slightly better performance than the existing approach.

• Management & Information Systems Review
• /
• v.32 no.3
• /
• pp.83-105
• /
• 2013

There are many literatures about the herding behavior of institutional investors but there is lack of literatures about the relation among several investor groups consisting of institutional investors. So we investigate the relation among sub-institutional investor groups like bank, insurance companies, pension funds using KRX intraday trading data of 2009. As the result, we find that foreign, individual, and securities firm investors trade in the opposite direction of other investor groups including pension funds. And pension, insurance, asset management, private equity funds, other companies, government, and banks are cross-mimicking each other, so we conclude that these investors make herding behavior. In 2009 institutional investors except securities firms make herding in a short period, and insurance, asset management, pension funds and other companies make herding and self-mimicking in all period, but there is no herding and mimicking after foreign investors.

• Proceedings of the KIEE Conference
• /
• 2009.07a
• /
• pp.1833_1834
• /
• 2009

본 논문에서는 이동 로봇의 군집 제어를 위해 선도-추종 로봇 방법에 기반한 적응 군집 제어 방법을 제안한다. 제안된 제어 방법 에서는 선도 로봇의 선형 속도와 각 속도를 추정하기 위해 투영 법칙에 기반한 적응 기법을 이용하며, 또한 백스테핑 제어 기법을 이용하여 이동 로봇의 제어 입력인 토크를 구한다. 마지막으로, 컴퓨터 모의실험을 통해 본 논문에서 제안한 제어기의 성능을 검증하고자 한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2010.11a
• /
• pp.527-530
• /
• 2010

효과적으로 군집 로봇을 제어하기 위해 리더 로봇을 선정하여 특별한 역할을 부여하는 방법이 보편적으로 사용되고 있다. 본 논문에서는 효과적인 지형 탐사를 위해 군집 로봇 그룹의 리더를 관리하는 방법을 제안한다. 그룹에 리더가 없는 경우와 리더를 선정한 경우로 나누어 2차원 지형을 탐사하는 실험을 수행하였다. 지형을 탐사하는 과정에서 리더 로봇이 동작 불능 상태가 되는 경우가 발생할 수도 있는데, 이때에는 그룹의 리더를 교체하는 방법을 사용하였다. 그룹의 리더를 관리하는 방법을 시뮬레이션 환경에 적용하여 실험한 결과, 리더가 없는 경우보다 지형 탐사 성공률이 훨씬 높아지는 것을 알 수 있었다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2011.04a
• /
• pp.275-278
• /
• 2011

체인로봇은 단일 로봇보다 강인성 측면에서 유리하기 때문에 무인탐사 환경에 적합하다. 체인로봇을 관리하기 위해서는 선두 로봇을 선정하고 군집을 관리하는 방법이 사용된다. 본 논문에서는 군집로봇의 효과적인 미로 탐색을 위해 체인로봇 그룹의 순위를 미로 환경에 맞게 적응적으로 순위를 재지정 하는 방법인 적응적 체인로봇 알고리즘을 제안한다. 그리고 체인 알고리즘을 적용한 경우와 적응적 체인 알고리즘을 적용한 경우로 나누어 2차원 맵을 탐색하는 실험을 수행하였다. 군집로봇이 맵을 이동하는 과정에서 선두 로봇이 동작 불능 상태가 되는 경우가 발생할 수도 있다. 이때 체인로봇을 적응적으로 순위를 재지정하는 방법을 사용하였다. 본 논문에서 제시한 방법을 시뮬레이션 환경에서 실험하였는데 실험 결과에 따르면 100%의 성공률을 얻을 수 있었다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2016.10a
• /
• pp.172-175
• /
• 2016

본 논문은 손목에 부착된 단일 3축 가속도 센서를 이용하여 사용자 움직임 기반의 휴식, 걷기, 달리기(느린속도, 빠른속도)를 분류하는 방법에 관한 연구이다. 초당 32회 표본 값의 가속도 정보에서 특징 신호인 평균, 표준편차를 산출하고 사용자의 행동상태를 4가지 상태로 분류한다. 분류 기준이 모호한 상태전이 신호에 대해 6가지 상태로 분류하여 구해진 총 10개의 행동상태 정보를 2차원 평면에 사영하고 최종적으로 K-means 군집화 기법을 적용하여 사용자의 행동상태를 4가지 상태로 분류한다.

• Korean Journal of Fisheries and Aquatic Sciences
• /
• v.30 no.5
• /
• pp.882-895
• /
• 1997

Changes in the benthic communities have been studied to investigate the environmental effects before and after the construction of Shihwa Dike in the West coast of Korea. It is suggested that sequential changes in macrofaunal assemblages progressed in two sucressional directions. In the Shihwa lake under the influence of organic enrichment. First, the appearance of 'azoic tone' or 'grossly polluted zone' developed in the area of less than 6 m in depth resulted from the severe dissolved oxygen depletion due to the eutrophication from the increased organic loading. Second, the 'polluted zone' characterized by the proliferation of the opportunistic species in organically enriched area, was found in the vicinity of the industrial discharges and nearby fluvial inputs. This benthic community succession in the Shihwa lake seemed to be caused by the various ecological events such as an eutrophication in this organically enriched environment after construction of the dike and other physico-chemical parameters like salinity and dissolved oxygen in the bottom water, which may be influenced by the irregular surface water discharge and dilution by outer seawater inflow through the water gate of the dike. On the other hand, the benthic communities in the outside of the dike showed that the species richness was more than doubled and the abundance increased almost seven times more than that before the dike construction. This may be a typical characteristics of the initial phase in benthic eutrophication, suggesting that an increased organic input area may have been reponsible for this faunal change in the study area.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.36 no.6
• /
• pp.565-573
• /
• 2008

The behavior-based decentralized approach is considered for multi-UAV formation flight. It is assumed that each UAV has its own mission of flying to a specified region, while the distances between UAVs should be maintained. These two requirements may conflict with each other. To design the controller, coupled dynamics approach is applied to multi-UAVs with an assumption that each UAV can communicate with each other to share the state-information. Control gain matrices are optimized to acquire better performances of formation flying. To validate the proposed control approach, numerical simulation is performed for the waypoint-passing mission of multi-UAVs.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.20 no.3
• /
• pp.441-446
• /
• 2010

Each robot decides and behaviors themselves surrounding circumstances in the swarm robot system. Robots have to conduct tasks allowed through cooperation with other robots. Therefore each robot should have the ability to learn and evolve in order to adapt to a changing environment. In this paper, we proposed learning based on Q-learning algorithm and evolutionary using Harmony Search algorithm and are trying to improve the accuracy using Harmony Search Algorithm, not the Genetic Algorithm. We verify that swarm robot has improved the ability to perform the task.

• Korean Journal of Environment and Ecology
• /
• v.26 no.3
• /
• pp.412-425
• /
• 2012

This study investigated the actual vegetation and plant community structure of Geumsun valley and Weonjeok valley of Naejangsan(Mt.) National Park to provide the data for park management. As a result of analysis of actual vegetation, Quercus variabilis community(47.3%) and Quercus serrata community(17.0%) are widely distributed but Carpinus tschonoskii community(7.5%) and Zelkova serrata community(6.9%) are distributed in valley forests. Plant communities were divided into 6 communities of Fraxinus mandshurica, Carpinus tschonoskii, Carpinus tschonoskii-Platycarya strobilacea-Zelkova serrata, Zelkova serrata, deciduous broad-leaved mixture(Zelkova serrata-Lindera erythrocarpa-Acer palmatum), and Lindera erythrocarpa-Cornus walteri according to the analysis of TWINSPAN of classification using 20 plots($400m^2$). Geumseon valley is composed of boulder stone area of average slope $20^{\circ}$, and there were distributed Fraxinus mandshurica community of 86 years old, Carpinus tschonoskii community in age from 56 to 79, and Carpinus tschonoskii-Platycarya strobilacea-Zelkova serrata community in age from 48 to 71. Weonjeok valley is composed of boulder stone area of average slope $11^{\circ}$, and there were distributed Zelkova serrata community in age from 52 to 71, deciduous broad-leaved mixture community in age from 49 to 70, and Lindera erythrocarpa-Cornus walteri community in age from 43 to 51. Ecological succession of each community was predicted to maintain same state. The Shannon's species diversity showed from 0.8220(Fraxinus mandshurica community) to 1.3850(Carpinus tschonoskii community) per unit area of $400m^2$.