• 한국음향학회지
• /
• 제31권1호
• /
• pp.1-10
• /
• 2012

수중음향채널에서 시변 다중경로 전달특성에 의해 발생하는 시변 인접심볼간 간섭을 완화할 수 있는 방법으로 선형 등화와 결정궤환 등화 기법을 사용할 수 있다. 그러나 선형 등화 기법은 간섭성분을 완벽하게 제거하지 못하는 단점이 있고, 결정궤환 등화기는 검출 오류 발생 시 오류 전파가 발생하는 문제를 가지고 있다. 본 논문에서는 이러한 문제점을 개선하기 위해 훈련 시퀀스를 이용하여 등화 모드를 선택하는 기법을 제안한다. 제안된 기법은 훈련시퀀스로부터 산출된 신호 대 잡음비에 따라 등화기법을 선택하여 동작시키는 방법으로 낮은 SNR일 때에는 선형 등화 기법이 적용하여 오류전파를 줄이고, 높은 SNR일 때에는 결정궤환 등화 기법을 적용하여 간섭성분을 제거함으로써 하나의 등화기법을 사용하는 경우보다 오류 성능을 향상시킬 수 있다. 해상실험을 통해 얻은 채널 응답을 이용하여 전산모의실험을 수행한 결과 제안된 기법이 효과적으로 오류 성능이 개선됨을 보였다.

• 한국통신학회논문지
• /
• 제31권11A호
• /
• pp.1167-1175
• /
• 2006

ML-DFE(Maximum Likelihood-Decision Feedback Equalization) 기법은 V-BLAST와 같은 공간다중화 MIMO시스템에서 ML 기법의 구현 복잡도를 줄이기 위한 준 최적 신호검출기법으로 불 수 있다. ML-DFE 기법은 ML 기법과 DFE 기법을 결합하여 오차전파를 줄이면서 rich scattering 환경에서 높은 다이버시티 이득을 얻을 수 있다. 본 논문에서는 ML-DFE 기법과 동일한 성능을 보이면서 연산복잡도를 줄일 수 있는 MML-DFE(Modified Maximum Likelihood - Decision Feedback Equalization) 기법을 제안한다. 또한 FPGA 구현을 통하여 제안된 MML-DFE 기법이 기존 ML-DFE 기법에 비하여 구현복잡도를 크게 감소시키면서 동일한 성능을 유지함을 확인한다.

• 대한전자공학회논문지TC
• /
• 제39권4호
• /
• pp.173-179
• /
• 2002

본 논문은 DFE (Decision Feedback Equalizer)구조와 CMA (Constant Modulus Algorithm), 그리고 LMS (Least Mean Square) 알고리즘을 이용한 등화기에 대하여 기술한다. DFE 구조는 기존의 transversal 구조의 등화기에 비하여 빠른 채널 적응 속도와 낮은 BER (Bit Error Rate) 값을 가지며 ISI(Intersymbol Interference)가 심한 환경에서도 좋은 성능을 나타낸다. 본 등화기는 16/64 QAM(Quadrature Amplitude Modulation) 변복조 방식에 적용할 수 있으며, 고속으로 동작할 수 있도록 고속의 곱셈기와 많은 수의 CSA (Carry Save Adder)를 사용하였다. COSSAP/sup TM/ 캐드 툴을 사용하여 부동 소수점 모델과 고정 소수점 모델을 개발하였으며, VHDL 모델을 개발하였다. 시뮬레이션 결과에 따라 feedback 부분과 feedforward 부분에 각각 12개와 8개의 탭을 사용하였으며, 다중 경로 페이딩 채널에서 BER이 10-6일 때를 기준으로 보면 등화기를 사용하지 않은 채널의 BER 보다 SNR(Signal to Noise Ratio)이 4dB 정도 향상되었다. SYNOPSYS/sup TM/ 캐드 툴과 삼성의 0.5 ㎛ standard cell library (STD80) 를 이용하여 로직 합성을 수행하였으며, 전체 게이트 카운트는 약 13만개를 보였다.

• Journal of Electrical Engineering and information Science
• /
• 제3권1호
• /
• pp.126-130
• /
• 1998

In this paper, a new interference rejection filter using the data-recycling technique is presented. The reference signal of the interference rejection filter is formed by using chip decisions, which is correlated with the narrowband interference components of the received signal. The decision feedback techniques reduce the distortion of the desired signal, which is introduced by the interference rejection filter through he use of feedback chip decisions. In order to update the filter coefficients a simple and efficient data-recycling technique that has improved performance over the conventional LMS algorithm is used. The performance of the proposed interference rejection filter is compared to filter with conventional LMS linear filter. The results show that the convergence speed of the proposed interference rejection filter using decision feedback and data-recycling technique is significantly faster than that of the conventional filters. Also, BER performance of the interference rejection filter demonstrates the superiority of the proposed filter algorithm.

• 한국통신학회논문지
• /
• 제21권10호
• /
• pp.2679-2686
• /
• 1996

In this paper, we propsoed the two methods to improve the conventional decision feedback interference canceller in DS/SS communication systems. The data bit is obtained by correlating the PN sequence with the received signals to the present time k, and thus the errors in the reference signal can be reduced by newly deciding all the reference signals with the resultant data bit. Additionally the cancelled signals are computed with less weight for initial reference signals of low processing gain, and highly weighted as the processing gain goes up. the resulting interference canceller outperforms the existing ones. By simulation, we found the proposed algorithm has "2-3 dB" performance gain at BER 10$^{-3}$ compared to the conventional descision feedback algorithm.algorithm.

• 한국시스템다이내믹스연구
• /
• 제16권2호
• /
• pp.5-32
• /
• 2015

Last year, a disastrous accident happened on the sea. Car ferry Se-wol which was on its way from In-cheon to Je-ju sank in the middle of the western sea. 304 passengers were drowned at the sea in the accident. There were no passengers who were waiting on board ship rescued in the accident. This study is to review the pattern of decision making in the middle of saving a life, right after Se-wol was about to sink. The starting point of this study is decision making on the rescue by rescue corps with no expert knowledge at the initial stage. When it comes to saving victims, there was no expert on rescue on the sea. Failure in rescue at Se-wol disaster lets us find a vicious circle of positive feedback loop. Experts with no experience act like well-qualified persons. Unqualified persons(leaders) have no expertise. Those who pretend to be expert occupy the important post in the hierarchy. It means we could not accumulate expertise of know-how in rescue division. That's why a vicious circle of positive feedback is formed in areas of disasters. It is necessary to train experts in Government Organizations to weaken the vicious circle of positive feedback.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2005년도 ICCAS
• /
• pp.1132-1136
• /
• 2005

In this paper, adaptive equalizer receiver for a turbo code direct sequence code division multiple access (DSCDMA) by using least mean square (LMS) adaptive algorithm is presented. The proposed adaptive equalizer is using soft output of decision feedback adaptive equalizer (DFE) to examines the output of the equalizer and the Log- maximum a posteriori (Log-MAP) algorithm for the turbo decoding process of the system. The objective of the proposed equalizer is to minimize the bit error rate (BER) of the data due to the disturbances of noise and intersymbol interference (ISI)phenomenon on the channel of the DS-CDMA digital communication system. The computer program simulation results shown that the proposed soft output decision feedback adaptive equalizer provides a good BER than the others one such as conventional adaptive equalizer, infinite impulse response adaptive equalizer.

• 한국정보시스템학회지:정보시스템연구
• /
• 제18권3호
• /
• pp.135-154
• /
• 2009

As the knowledge is recognized as a core factor of organization's competitiveness and creation of value added, the importance of knowledge management is also increased. To achieve the successful knowledge management, it is important to establish strategy that consider essential purpose of knowledge management such as creating and sharing of knowledge resource, improving performance, and continuing organizational innovation within the organization and influence factor inside and outside of organization. Until now, however, the research for knowledge management strategy was mostly limited to the statistical analysis based on the unilinear causality model, and systematic access and analysis that consider interaction and feedback structure between factors. In this paper, we developed the novel decision-making framework for successful strategy establishment by applying the analytic network process(ANP). Specifically. we derive clusters and components to decide the interaction and feedback structure between the elements of knowledge management by literature studies. And we produced relative importance and preference of clusters, components and alternatives dealing with feedback structure through the survey of experts in the field or related one of knowledge management. In result of this study, we expect that it will help the knowledge officer to decide establishing knowledge management strategy.

• 반도체디스플레이기술학회지
• /
• 제14권2호
• /
• pp.9-16
• /
• 2015

In this paper, selection criterions on selection diversity are researched. The diversity is applied to the multiple antenna system based on wireless access in vehicular environment (WAVE) standard for rapid varying channel. Least squares (LS) based decision feedback equalizer (DFE) are used for channel equalization. Received signal is regenerated by means of the decision feedback path. In the selection diversity, the regenerated signal as well as the received signal is selected according to selection criterion. The decision feedback algorithm can follow the fast speed of WAVE fading channel. To control the tracking speed of the time-varying channel, simple low pass filter is used. Finally, the estimated channel value recovers the distorted payloads. Signal power before automatic gain control (AGC) in analog stage can be used as a selection criterion. In the digital stage, signal power after AGC, noise power after AGC, signal to noise ratio after AGC and cross-correlation method can be used as selection criterions. According to the simulation results, the performance of the selection diversity is improved in comparison with that of the combining diversity for the WAVE fading channel.

• 한국전자파학회지:전자파기술
• /
• 제5권3호
• /
• pp.22-35
• /
• 1994

본 논문에서는 대역폭 효율의 향상을 위해 상관 부호에 따른 FM시스템올 설계한다. 상판 부호화된 신호는 연속 위상과 정진폭 특성으로 전력 효율이 높다. 하지만 이 신호는 최소 유콜리언 거리(${d^2}_{mm}$)가 작기 때분에 신호대 잡음비(SNR)의 손실과 수신 성능(BER)의 저하를 초래한다. 따라서 본 논문에서는 수신 성능의 향상을 위하여 상관 부호화에 따른 수정된 최대 근사 추정을 작용한다. 최대 근사 추정은 사용된 채널에서 bit-by-bit 검파에 비하여 수산 오율을 향상시킨다. 결정 궤환을 적용하지 않은 경우, 최대 근사 추정은 bit by bit 겁파에 비해 약 2dB의 수신 성능이 향상되었다. 결정 궤환을 적용한 경우, 최대 근사 추정은 bit-by bit 검파에 비해 약 4dB의 수신 성능이 향상되었다. 특히 안정된 채널보다 채널이 열악할수록 최대 근사 추정과 결정 궤환의 성능이 우수하다.