• 한국결정성장학회지
• /
• 제8권1호
• /
• pp.42-48
• /
• 1998

PLT($Pb_{1-x}La_{2x/3}TiO_3$, x=0.1~0.3) 분말의 합성을 수열법에 의해 행하였다. 최적의 PLT분말을 얻을 수 있는 합성조건은 반응온도; $250^{\circ}C$, 수열용매; 8M-KOH, 반응시간; 12시간이었으며, $Pb_{1-x}La_{2x/3}TiO_3$의 조성이 $x{\leq}0.2$일 때 단일상의 PLT가 얻어졌다. 수열적으로 얻어진 PLT분말에 있어서 tetragonality(c/a)와 상전이 온도는 고용된 La2O3 함량의 증가에 따라 감소하였으며, $Pb_{1-x}La_{2x/3}TiO_3$(x=0.2)의 Curie 온도는 약 $400^{\circ}C$이었다. 합성분말의 형태는 구상에 가깝고, 입자의 크기는 20~200mm 범위였다.

• 한국지능정보시스템학회:학술대회논문집
• /
• 한국지능정보시스템학회 2001년도 The Pacific Aisan Confrence On Intelligent Systems 2001
• /
• pp.59-64
• /
• 2001

The robot soccer simulation game is a dynamic multi-agent environment. In this paper we suggest a new reinforcement learning approach to each agent's dynamic positioning in such dynamic environment. Reinforcement learning is the machine learning in which an agent learns from indirect, delayed reward an optimal policy to chose sequences of actions that produce the greatest cumulative reward. Therefore the reinforcement learning is different from supervised learning in the sense that there is no presentation of input pairs as training examples. Furthermore, model-free reinforcement learning algorithms like Q-learning do not require defining or learning any models of the surrounding environment. Nevertheless it can learn the optimal policy if the agent can visit every state- action pair infinitely. However, the biggest problem of monolithic reinforcement learning is that its straightforward applications do not successfully scale up to more complex environments due to the intractable large space of states. In order to address this problem. we suggest Adaptive Mediation-based Modular Q-Learning (AMMQL)as an improvement of the existing Modular Q-Learning (MQL). While simple modular Q-learning combines the results from each learning module in a fixed way, AMMQL combines them in a more flexible way by assigning different weight to each module according to its contribution to rewards. Therefore in addition to resolving the problem of large state effectively, AMMQL can show higher adaptability to environmental changes than pure MQL. This paper introduces the concept of AMMQL and presents details of its application into dynamic positioning of robot soccer agents.

• 한국해양공학회지
• /
• 제20권5호
• /
• pp.77-81
• /
• 2006

In this study, monolithic liquid phase sintered SiC (LPS-SiC) was made by the hot pressing method with nano-SiC powder, whose particle size is 30 nm and less on the average. Alumina ($Al_{2}O_{3}$), yttria ($Y_{2}O_{3}$), and silica ($S_{i}O_{2}$) were used for sintering additives. To investigate the effects of $S_{i}O_{2}$, the $Al_{2}O_{3}/Y_{2}O_{3}$ composition was fixed and the ratio of $S_{i}O_{2}$ was changed, with seven different ratios tested. And to investigate the effects of the sintering temperature, the sintering temperature was changed, with $1760^{\circ}C,\;1780_{\circ}C$, and $1800_{\circ}C$ being used with a $S_{i}O_{2}$ ratio of 3 wt%. The materials were sintered for 1 hour at $1760^{\circ}C,\;1780^{\circ}C$ and $1800^{\circ}C$ under a pressure of 20 MPa. The effects on sintering from the sintering system used, as well as from the composition of the sintering additives, were investigated by density measurements. Mechanical properties, such as flexural strength, were investigated to ensure the optimum conditions for a matrix of SiCf/SiC composites. Sintered densityand the flexural strength of fabricated LPS-SiC increased with an increase in sintering temperature. Particularly, the relative density of a sintered body at $1800^{\circ}C$ with a non-content of $S_{i}O_{2}$, a specimen of AYSO-1800, was 95%. Also, flexural strength was about 750MPa.

• 정보처리학회논문지D
• /
• 제11D권1호
• /
• pp.203-210
• /
• 2004

지리정보시스템에 대한 연구는 개방형 구조, 상호운영성과 확장 가능성에 초점을 두고 발전되어 왔다. 이 연구들은 기존에 구축된 하나의 통합시스템의 형태에서 대용량의 데이터, 고차원 구조의 데이터 그리고, 서로 다른 지리 포멧의 데이터를 처리하기 위하여, 응용 프로그램과 사용자 인터페이스를 레이어 별로 구분하여 각기 관리하는 개방형 구조로 시스템이 구현되어야 함을 의미한다. 현재 구현되는 개방형 지리정보시스템을 위한 미들웨어들 중 CORBA 미들웨어는 특정 운영체제를 기반으로 하지 않아 많은 개발자들이 이용하고 있다. 이때, CORBA의 서비스는 점대점 통신만을 지원하고, 다른 미들웨어보다 데이터 제공 시간이 많이 소요되는 문제점을 가지고 있다. 따라서 이 논문에서는 점대점 통신기법의 수정 방법을 제안하고, 구현한다. 제시된 방법은 클라이언트가 서버의 데이터를 한번의 연결로 데이터를 제공함으로써 그룹 통신과 빠른 데이터 제공 속도를 보장할 수 있게 한다.

• 한국실내디자인학회논문집
• /
• 제24권1호
• /
• pp.12-22
• /
• 2015

When it comes to the architecture design, Kuma Kengo rejects a strong and violent subject-centered position and advocates the design that is object-oriented. As can be seen in 'gentle architecture', 'three lows principle', 'natural architecture', and 'connecting architecture', he clearly expresses the objective nature of architecture design in those terms. In this respect, the purpose of this study is to make a close inquiry into the meaning, effect and characteristics of objectivity. In particular, we try to identify the contents of 'impure architecture', which has a clear ambivalence to be an instrumental expression strongly settled in the objectivity, in an aesthetic standpoint. To do that, we systemized the concept of mimesis and the theory of subversive appropriation by Walter Benjamin and Theodor Adorno in to a frame of interpretation. By systemizing translation, subversion, verbalization and the dialectic structure of the aesthetics of negation, we interpreted the features of his works as an objective work and 'impure architecture'. His objectivity leads the situation by subversively appropriating the inherent elements of architectural conditions based on a dialectic solution in which inquiries on logical and scientific materials have played a critical role. Above all, through all these processes, he tried to suggest a language as a new technique for materials and structures. Ultimately, we could find out that this object oriented design sublates a subject oriented way that is monolithic and repetitive regardless of objects. Rather, it is a way that is effective in creating a new way of design by making a different approach to a new object rather unfamiliarly, yet deeply.

• 해양환경안전학회지
• /
• 제22권2호
• /
• pp.181-186
• /
• 2016

본 연구에서는 해양 수산분야의 대학생 자원봉사활동에 대한 자료를 수집하여 활동유형을 유류방제작업, 어촌봉사활동, 해안정화작업 등으로 분석하였다. 또한 자원봉사 참여자 대부분은 환경보호, 보람, 사회교류, 특별체험, 자연과의 동화측면에서 만족하고 있었다. 주제측면에서 체계성, 효율성, 활동의 다양성, 자원홍보활용, 봉사이후 보상, 자기봉사역량 측면에서 불만족스러워 한다. 이러한 연구 결과를 바탕으로 대학생 자원봉사의 정책적 측면에서 해양 수산분야의 자원봉사를 증진시킬 수 있는 방안을 제안하였다.

• Ocean Systems Engineering
• /
• 제7권3호
• /
• pp.179-193
• /
• 2017

Dynamic response analysis of offshore triceratops with stiffened buoyant legs under impact and non-impact waves is presented. Triceratops is relatively new-generation complaint platform being explored in the recent past for its suitability in ultra-deep waters. Buoyant legs support the deck through ball joints, which partially isolate the deck by not transferring rotation from legs to the deck. Buoyant legs are interconnected using equally spaced stiffeners, inducing more integral action in dispersing the encountered wave loads. Two typical nonlinear waves under very high sea state are used to simulate impact and non-impact waves. Parameters of JONSWAP spectrum are chosen to produce waves with high vertical and horizontal asymmetries. Impact waves are simulated by steep, front asymmetric waves while non-impact waves are simulated using Stokes nonlinear irregular waves. Based on the numerical analyses presented, it is seen that the platform experiences both steady state (springing) and transient response (ringing) of high amplitudes. Response of the deck shows significant reduction in rotational degrees-of-freedom due to isolation offered by ball joints. Weak-asymmetric waves, resulting in non-impact waves cause steady state response. Beat phenomenon is noticed in almost all degrees-of-freedom but values in sway, roll and yaw are considerably low as angle of incidence is zero degrees. Impact waves cause response in higher frequencies; bursting nature of pitch response is a clear manifestation of the effect of impact waves on buoyant legs. Non-impact waves cause response similar to that of a beating phenomenon in all active degrees-of-freedom, which otherwise would not be present under normal loading. Power spectral density plots show energy content of response for a wide bandwidth of frequencies, indicating an alarming behaviour apart from being highly nonlinear. Heave, being one of the stiff degrees-of-freedom is triggered under non-impact waves, which resulted in tether tension variation under non-impact waves as well. Reduced deck response aids functional requirements of triceratops even under impact and non-impact waves. Stiffened group of buoyant legs enable a monolithic behaviour, enhancing stiffness in vertical plane.

• 한국통신학회논문지
• /
• 제32권5A호
• /
• pp.497-502
• /
• 2007

피드포워드 방식을 이용하여 광대역, 선형 이득 제어 특성을 갖는 SiGe HBT 가변 이득 증폭기를 설계 및 제작하였다. 가변 이득 증폭기는 능동 발룬, 차동형 주 증폭기, 피드포워드 블록, 전압 조절부로 구성 되었으며, 주 증폭기와 피드포워드 블록의 신호가 역위상으로 상쇄되어 광대역의 선형 이득 제어가 가능하도록 각 부분을 최적화 시켰다. 설계된 가변 이득 증폭기는 STMicroelectronics사(社)의 0.35 ${\mu}m$ Si-BiCMOS 공정을 이용하여 제작하였다. 제작 및 측정 결과, 피드포워드 방식의 가변 이득 증폭기는 4 GHz($4\;GHZ{\sim}8\;GHz$)의 광대역 특성을 나타내었다. 또한, 제작된 가변 이득 증폭기는 6 GHz에서 9.3 dB의 최대 이득과 0.6 - 2.6 V의 조절 전압 인가시 19.6 dB의 이득 조절 범위 특성을 나타내었으며, 8 GHz에서 -3 dBm의 출력 전력 특성을 각각 나타내었다.

• 한국전기전자재료학회:학술대회논문집
• /
• 한국전기전자재료학회 1997년도 춘계학술대회 논문집
• /
• pp.135-140
• /
• 1997

The influences of ZrB$_2$ additions to SiC on microstructural, DDM(Electrical Discharge Machining), mechanical and electrical properties were investigated. composites were prepared by adding 15, 30, 45 vol.% ZrB$_2$particles as a second phase to SiC matrix. SiC-ZrB$_2$ composites obtained by hot pressing for high temperature structural application were fully dense with the relative densities over 99%. The fracture toughness of the composites were increased with the ZrB$_2$contents. In case of composite containing 30vol.% ZrB$_2$, the flexural strength and fracture toughness showed 45% and 60% increase, respectively compared to that of monolithic SiC sample. The electrical resistivities of SiC-ZrB$_2$ composites decreased significantly with the ZrB$_2$ contents. The electrical resistivity of SiC-30vol.% ZrB$_2$ composite showed 6.50$\times$10$^{-4}$ $\Omega$.cm. Cutting velocity of EDM of SiC-ZrB$_2$ composites are directly proportional to duty factor of pulse width. Surface roughness, however, are not all proportional to pulse width. Higher-flexural strength composites show a trend toward smaller crater volumes, leaving a smoother surface; the average surface roughness of the SiC-ZrB$_2$ 15 vol.% composite with the flexural strengthe of 375㎫ was 3.2${\mu}{\textrm}{m}$, whereas the SiC-ZrB$_2$ 30.vol% composite of 457㎫ was 1.35${\mu}{\textrm}{m}$. In the SEM micrographs of the fracture surface of SiC-ZrB$_2$ composites, the SiC-ZrB$_2$ two phases are distinct; the white phase is the ZrB$_2$and the gray phase is the SiC matrix. In the SEM micrographs of the EDM surface, however, these phases are no longer distinct because of thicker recast layer of resolidified-melt-formation droplets present. It is shown that SiC-ZrB$_2$ composites are able to be machined without surface cracking.

• 한국시뮬레이션학회:학술대회논문집
• /
• 한국시뮬레이션학회 2001년도 The Seoul International Simulation Conference
• /
• pp.321-324
• /
• 2001

The robot soccer simulation game is a dynamic multi-agent environment. In this paper we suggest a new reinforcement learning approach to each agent's dynamic positioning in such dynamic environment. Reinforcement Beaming is the machine learning in which an agent learns from indirect, delayed reward an optimal policy to choose sequences of actions that produce the greatest cumulative reward. Therefore the reinforcement loaming is different from supervised teaming in the sense that there is no presentation of input-output pairs as training examples. Furthermore, model-free reinforcement loaming algorithms like Q-learning do not require defining or loaming any models of the surrounding environment. Nevertheless it can learn the optimal policy if the agent can visit every state-action pair infinitely. However, the biggest problem of monolithic reinforcement learning is that its straightforward applications do not successfully scale up to more complex environments due to the intractable large space of states. In order to address this problem, we suggest Adaptive Mediation-based Modular Q-Learning(AMMQL) as an improvement of the existing Modular Q-Learning(MQL). While simple modular Q-learning combines the results from each learning module in a fixed way, AMMQL combines them in a more flexible way by assigning different weight to each module according to its contribution to rewards. Therefore in addition to resolving the problem of large state space effectively, AMMQL can show higher adaptability to environmental changes than pure MQL. This paper introduces the concept of AMMQL and presents details of its application into dynamic positioning of robot soccer agents.