• 한국산학기술학회논문지
• /
• 제17권8호
• /
• pp.17-22
• /
• 2016

고차 QAM 시스템에 대한 자기적응 등화에서 눈 모형이 완전히 닫힌 등화 초기에 적용하여 눈 모형을 빠르게 열뿐만 아니라 정상상태 오차 레벨을 크게 낮추는 새로운 SDD 알고리즘을 제안한다. 제안 방법은 M-QAM 응용에서, 관찰에 가장 인접한 두 심볼을 추정의 기반으로 함으로써 기존 SDD의 계산 복잡성을 최소화하고, QAM 차수에 무관하게 연판정을 크게 단순화하였다. 아울러 심볼 추정에 임계 함수에 비해 오판정 회피가 우수한 시그모이드 함수를 적용, 추정의 신뢰도를 높였다. 또한 등화기 갱신을 위한 오차 발생 시 임계 함수에 의한 심볼 판정 값을 오차에 가중하여 오차 변동 범위를 확장함으로써 제안한 자기적응 등화기의 초기화 성능을 개선하였다. 결과적으로 제안 방법은 기존 SDD의 계산 복잡성과 초기화 및 수렴 특성을 현저히 개선하였다. 부가 잡음이 존재하는 다중경로 채널 조건에서 64-QAM 및 256-QAM에 대한 모의실험을 통해 CMA와 제안한 2-SDD 및 가중된 2-SDD의 두 가지 형태의 성능을 비교하고 제안 방법의 유용성을 확인하였다.

• 사물인터넷융복합논문지
• /
• 제9권3호
• /
• pp.35-43
• /
• 2023

본 연구는 AI 기반 맞춤형 학습 시스템을 시범적으로 운영하여 대학 수업에서의 AI 기반 맞춤형 학습 시스템의 적용 가능성과 효용성을 알아보고자 하였다. 이를 위하여 C지역 소재 B대학교 1학년 재학생 중 기초수학 교과목 수업에 참여한 42명 학습자를 대상으로 AI 기반 맞춤형 학습 시스템을 적용 및 운영하였고, 학생 및 교수를 대상으로 설문 문항 조사와 인터뷰를 진행하였다. 연구 결과, AI 기반 맞춤형 학습 시스템의 활용은 학생의 학업성취도를 향상시켰다. 심층인터뷰 결과 교수자와 학습자 모두 기초 개념 학습에 있어 학습 성과 향상에 기여하는 것으로 파악되었다. 이는 AI 기반의 맞춤형 학습 시스템이 자기 주도 학습의 역량을 향상하고 개념학습을 통해 지식 강화에 효과적인 방안이 될 것임을 시사한다. 본 연구는 인공지능 기반 적응형 학습 시스템의 기초 과학 교과목 도입과 적용에 관련한 기초자료로 활용될 수 있을 것이다. 향후 AI 기반 맞춤형 학습에서 학생들에게 제공한 학습과정과 분석한 데이터를 대면수업에 연계한 효과 검증과 분석한 데이터의 활용 방안에 대한 전략 연구를 제언한다.

• Nuclear Engineering and Technology
• /
• 제27권3호
• /
• pp.336-343
• /
• 1995

적응제어의 한 방식인 자기동조제어(STR) 방식이 비선형 노심 모델의 출력 조정에 적용된다. 적응제어는 비선형, 시변 및 확률(Stochastic) 시스템을 위한 준최적 제어기를 설계하기 위한 적절한 제어 방식이다. 제어계통은 미지의 시변 파라메타를 갖는 3차 선형 모델에 기초한다. 파라메타는 가변 망각계수를 도입한 늑장 최소자승법에 의하여 매시간(Time Step) 순환적으로 평가된다. 평가된 파라메타를 이용하여 한 스텝 먼저 냉자재 평균온도가 예측되고 이 예측된 값과 Setpoint 값과의 차이를 최소화함은 물론, 제어봉의 움직임을 막고자 가중(Weighted) One-step-ahead 제어기가 설계된다. 또한 적분동작이 첨가되어 정상상태 에러가 제거된다. 넓은 운전영역을 포괄하는 비선형 PWR 모델이 원자로 출력 조정을 위한 본 제어기를 시뮬레이션하는데 이용되었다. 시뮬레이션 결과로부터 본 제어기의 성능이 우수한 것으로 판명되었다.

• Journal of Biosystems Engineering
• /
• 제39권3호
• /
• pp.174-183
• /
• 2014

Purpose: A robust, efficient auto-grading computer vision system for meat carcasses is in high demand by researchers all over the world. In this paper, we discuss our study, in which we developed a system to speed up line processing and provide reliable results for pork grading, comparing the results of our algorithms with visual human subjectivity measurements. Methods: We differentiated fat and lean using an entropic correlation algorithm. We also developed a self-designed robust segmentation algorithm that successfully segmented several porkcut samples; this algorithm can help to eliminate the current issues associated with autothresholding. Results: In this study, we carefully considered the key step of autoextracting lean tissue. We introduced a self-proposed scheme and implemented it in over 200 pork-cut samples. The accuracy and computation time were acceptable, showing excellent potential for use in online commercial systems. Conclusions: This paper summarizes the main results reported in recent application studies, which include modifying and smoothing the lean area of pork-cut sections of commercial fresh pork by human experts for an auto-grading process. The developed algorithms were implemented in a prototype mobile processing unit, which can be implemented at the pork processing site.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제16권8호
• /
• pp.2787-2800
• /
• 2022

The aging society increases emergency situations of the elderly living alone and a variety of social crimes. In order to prevent them, techniques to detect emergency situations through voice are actively researched. This study proposes CutPaste-based anomaly detection model using multi-scale feature extraction in time series streaming data. In the proposed method, an audio file is converted into a spectrogram. In this way, it is possible to use an algorithm for image data, such as CNN. After that, mutli-scale feature extraction is applied. Three images drawn from Adaptive Pooling layer that has different-sized kernels are merged. In consideration of various types of anomaly, including point anomaly, contextual anomaly, and collective anomaly, the limitations of a conventional anomaly model are improved. Finally, CutPaste-based anomaly detection is conducted. Since the model is trained through self-supervised learning, it is possible to detect a diversity of emergency situations as anomaly without labeling. Therefore, the proposed model overcomes the limitations of a conventional model that classifies only labelled emergency situations. Also, the proposed model is evaluated to have better performance than a conventional anomaly detection model.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1996년도 Proceedings of the Korea Automatic Control Conference, 11th (KACC); Pohang, Korea; 24-26 Oct. 1996
• /
• pp.24.1-24
• /
• 1996

The development of future aircraft that involves the expanded flight envelop will place increased performance requirements on the design of the flight control system. Maneuvering areas are expanding into flight envelopes characterized by significantly larger levels of modeling uncertainty than encountered in present flight control designs. Conventional flight control techniques that ignore the effects of large parameter variations, modeling uncertainties and nonlinearities, will likely produce designs with poor performance and robustness. Recent advances in modern control theories called advanced control theories, most notably the H$\_$.inf./ synthesis technique, adaptive control and neural network application, offer the promise of a design technique that can produce both high performance and robust controllers for next generation aircraft. This special lecture will survey the recent development in advanced flight control and review the possible application of advanced control theories.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2003년도 ICCAS
• /
• pp.2278-2282
• /
• 2003

In this paper the structure of 2-DOF controller based on artificial immune network algorithms has been suggested for nonlinear system. Up to present time, a number of structures of the 2-DOF controllers are considered as 2-DOF (2-Degrees Of Freedom) control functions. However, a general view is provided that they are the special cases of either the state feedback or the modification of PID controllers. On the other hand, the immune network system possesses a self organizing and distributed memory, also it has an adaptive function by feed back law to its external environment and allows a PDP (parallel distributed processing) network to complete patterns against the environmental situation, since antibody recognizes specific antigens which are the foreign substances that invade living creatures. Therefore, it can provide optimal solution to external environment. Simulation results by immune based 2-DOF controller reveal that immune algorithm is an effective approach to search for 2-DOF controller.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2003년도 ICCAS
• /
• pp.1531-1536
• /
• 2003

In this paper the structure of 2-DOF controller based on artificial immune network algorithms has been suggested for nonlinear system. Up to present time, a number of structures of the 2-DOF controllers are considered as 2-DOF (2-Degrees Of Freedom) control functions. However, A general view is provided that they are the special cases of either the state feedback or the modification of PID controllers. On the other hand, The immune network system possesses a self organizing and distributed memory, also it has an adaptive function by feed back law to its external environment and allows a PDP (parallel distributed processing) network to complete patterns against the environmental situation, since antibody recognizes specific antigens which are the foreign substances that invade living creatures. Therefore, it can provide optimal solution to external environment. Simulation results by immune based 2-DOF controller reveal that immune algorithm is an effective approach to search for 2-DOF controller.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1991년도 한국자동제어학술회의논문집(국제학술편); KOEX, Seoul; 22-24 Oct. 1991
• /
• pp.1570-1575
• /
• 1991

The feasibility of using an adaptive resonance network (ART2) with unsupervised learning capability for too] wear detection in turning operations is investigated. Specifically, acoustic emission (AE) and cutting force signals were measured during machining, the multichannel AR coefficients of the two signals were calculated and then presented to the network to make a decision on tool wear. If the presented features are significantly different from previously learned patterns associated with a fresh tool, the network will recognize the difference and form a new category m worn tool. The experimental results show that tool wear can be effectively detected with or without minimum prior training using the self-organization property of the ART2 network.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제10권1호
• /
• pp.238-256
• /
• 2016

Modern web services have been broadly deployed on the Internet. Most of these services use multi-tier architecture for flexible scaling and software reusability. However, managing the performance of multi-tier web services under dynamic and unpredictable workload, and different resource demands in each tier is a critical problem for a service provider. When offering quality of service assurance with least resource usage costs, web service providers should adopt self-adaptive resource provisioning in each tier. Recently, a number of rule- and model-based approaches have been designed for dynamic resource management in virtualized data centers. This survey investigates the challenges of resource provisioning and provides a competing assessment on the existing approaches. After the evaluation of their benefits and drawbacks, the new research direction to improve the efficiency of resource management and recommendations are introduced.