http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
이상부(Sang-Boo Lee) 한국정보기술학회 2010 한국정보기술학회논문지 Vol.8 No.11
The researches that remote control systems and the monitoring data for long/short distances are continued using the network connections. Also sending/receiving as well as monitoring the various communication protocols are needed in various industries. Too much time and man power are needed to develop these communication protocols. To save this time and man power, in this paper, we reported the novel monitoring system which is transmitting/receiving and monitoring the data in one system. This system is integrated the TCP/IP Ethernet communication; ZigBee communication for sensor networking; RS485 for great distance and RS232 for short distance serial communications. For realizing the trustworthy total monitoring system, four kinds of communication protocol interfacing application program are developed and used. According to the experient result, total monitoring system had accurately received the data transferred through various communication protocol from end devices, and showed its superb performance.
상황정보 무선통신을 위한 ZigBee 응용시스템 구현에 관한 연구
이상부(Sang-Boo Lee) 한국정보기술학회 2009 한국정보기술학회논문지 Vol.7 No.5
The wire communication system has a high accuracy, but it needs higher expense for its establishment and has a more complex connection than radio communication system. And it is complicated to add and remove sensor nodes to the wire communication system. For these reasons, using ZigBee radio communication technology is being tried in many fields to radio monitoring or radio remote control. The ZigBee communication consumes a low power, so its battery using time is much longer and more efficient relative to the existent radio communication or bluetooth communication system. And many sensor nodes can be installed to the network due to low expenses for system implementation. A Sink node receives data from many router sensor nodes. It is needed to develop the ZigBee application system to remote control and remote monitor the devices by data of sensor node. The ZigBee application system which is applied to remote sensor network and useful for radio remote control and monitor will be implemented in this paper.
퍼지 제어기와 퍼지 신경망제어기의 응답 특성에 관한 연구
김형수(Kim Hyung Soo),이상부(Lee Sang Boo),김홍기(Kim Hong Gi) 한국정보처리학회 1996 정보처리학회논문지 Vol.3 No.6
this study examines the response characteristics of the fuzzy controller and the fuzzy neural network controller. The former is excellent in terms of the overshoot at its initial values and has great advantages on the disturbance. But there exist some errors in its desired output. Many methods have been introduced that remove the errors of the desired state. This study is in more favor of the fuzzy neural network controller using the neural network than any other method. The fuzzy neural network controller complements the shortcomings of fuzzy controller and can be an accurate controller by being well-without any disturbance ir error-converged to the desired output. And it is through simulation that the comparison of the two controllers is carried out in this study.
이권윤(Gwon-Yoon Lee),이상부(Sang-Boo Lee) 한국정보기술학회 2012 한국정보기술학회논문지 Vol.10 No.12
Researches have been conducted continuously in order to communicate both short and long distance information with an integrated communication system by using diverse systems such as wire and wireless communication. In various industry fields, it is required to communicate and control by using communication protocols. In this paper, we developed an integrated communication system which can communicate RS232 communication which is appropriate for a short distance serial communication, RS485 which is commonly used for long distance serial communication, TCP/IP for the Internet communication, Zigbee radio communication mainly used for sensor networking system, and Bluetooth which is related to a short distance radio communication with using one system. For a trustworthy integrated communication system, we developed an application program which is able to interface five kinds of communication protocols. According to the experimental results, the integrated communication system accurately sent, received and transferred data through the five communication protocols in real time and showed superb performance.
이권윤(Gwon-Yoon Lee),이상부(Sang-Boo Lee) 한국정보기술학회 2018 한국정보기술학회논문지 Vol.16 No.10
Various algorithms and models of deep learning were developed, and RNN shows good performance in future prediction system. RNN and LSTM applied universal prediction system was realized in this paper, and performance was evaluated by cost value. Difference between input data was so big that we normalized this input data, and sequence size of learning input was 5 in the performance experiment. The legibility of the results was increased by displaying learning result, predictive execution result and cost values on a real-time graph. In this experiment, accomplishment of learning was confirmed by cost values decrease from learning number 150 to 0.085. future prediction system precisely learned learnin target value and test target value in a small number of learning time in a short time. Furthermore, cost value decreased to wanted value, confirming that the performance of suggested prediction system is excellent.
김형수(Hyung Soo Kim),김홍기(Hong Gi Kim),이상부(Sang Boo Lee) 한국정보과학회 1998 정보과학회논문지(B) Vol.25 No.9
대용량의 정보시스템에서 유용한 정보를 검색하여 효율적인 의사결정을 하기 위해서는 정제된 지식베이스의 구축과 정형화된 검색 연산이 필요하다. 러프집합이론에 있어서 다중 속성리덕트의 생성과 퍼지 질의에 대한 경계역 객체의 분류의 문제점을 개선하고 자, 본 연구에서는 결정 및 조건속성에 대한 상관분석과 베이지안 정리를 적용한 근사 결정규칙을 생성하는 알고리듬을 제안한다. 제안된 알고리듬에 의해 생성된 최소 지식베이스에서 소속함수 및 t-노름의 퍼지연산을 정의하여 퍼지 정보를 검색하기 위한 근사적 추론 절차를 기술한다. 또한 본 연구의 적용사례로써 서비스 요인을 고려한 호텔결정시스템 HSRF을 구현한다. 임의의 객체와 속성요인의 도메인을 난수 발생시키는 시뮬레이션을 통하여 본 연구의 러프·퍼지 검색모델과 기존 검색이론과의 비교 결과, 본 연구가 데이타베이스 크기에 따른 객체검색에 있어 탐색시간의 효율성이 있음을 보인다. In order to make efficient decision to retrieve useful information in the large information system, it is necessary to implement refined knowledge base and operate formatted retrieval. Rough set has difficulty in generating multiple reducts and classifying boundary objects for fuzzy query. In this paper, we propose algorithms generating approximate decision rule by correlation analysis and Bayesian theorem to improve these restrictions. We define membership function and t-norm operation in minimal knowledge base created by these algorithms, and describe procedure of approximate inference to retrieve fuzzy information. As an applied example of this paper, we also implement the HSRF in consideration of hotel service factors. Through simulation randomly generating object number & attribute domain, we compare our retrieval model based on rough & fuzzy theory with the other retrieval theories. As the result, our modeling method is more efficient than the previous theories in access time for object retrieval according to database size.