• Proceedings of the Korean Information Science Society Conference
• /
• 2011.06c
• /
• pp.331-334
• /
• 2011

입력 데이터의 공통적인 특징을 찾아내는 방법은 기계 학습 분야의 중요한 분야이다. 일반적으로 입력 데이터의 형태적 패턴을 찾아내는 알고리즘들이 많이 연구되었는데, 최근에는 데이터의 입력 순서 또는 데이터 사이의 시간적 인과 관계와 같이 시간에 연관된 패턴을 찾는 방법이 주목을 받고 있다. 우리는 형태적 혹은 공간적 패턴 탐색에 뛰어난 성능을 보이는 하이퍼네트워크 모델을 확장하여 입력 데이터의 시공간적 패턴을 찾는 방법을 제시한다. 하이퍼네트워크는 두 개 이상의 변수를 하나의 엣지로 연결하여 문제공간을 탐색하는 모델로, 시간과 공간의 변수를 동시에 고려하여 데이터의 특성을 찾아내는 데에 적합하다. 이를 확인하기 위하여 사람의 EEG 신호를 분석하였는데, 시각적인 정보를 처리할 때와 언어적 정보를 처리할 때의 특징적인 패턴들을 찾았다.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.2C
• /
• pp.222-228
• /
• 2004

This paper describes a new adaptive algorithm which can reduce the required computation quantities in the adaptive filter. The proposed adaptive algorithm uses only the signs of the normalized input signal rather than the input signals when coefficients of the filter are adapted. By doing so, there is no need for the multiplications and divisions which are mostly responsible for the computation quantities. To analyze the convergence characteristics of the proposed algorithm, the condition and speed of the convergence are derived mathematically. Also, we simulate an echo canceller adopting this algorithm and compare the performances of convergence for this algorithm with the ones for the other algorithm. As the results of simulations, it is proved that the echo canceller adopting this algorithm shows almost the same performances of convergence as the echo canceller adopting the SIA algorithm.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2002.05a
• /
• pp.264-267
• /
• 2002

This paper presents the design of a pipelined virtual output queue routing engine for an advanced input-queued ATM switch, which has a serial cross bar structure. The proposed routing engine has been designed for wire-speed routing with a pipelined buffer management. It provides the tolerance of requests and grants data transmission latency between the routing engine and central arbiter using a new request control method that is based on a high-speed shifter. The designed routing engine has been implemented in a field programmable gate array (FPGA) chip with a 77MHz operating frequency, 16$\times$16 switch size, and 2.5Gbps/port speed.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.12 no.1
• /
• pp.85-92
• /
• 1998

In this paper, we design a sequential system controller, which is capable of processing parallel sequence, on the basis of analysis of control specification described by Petri Net with incidence matrix. The sequential system controller consists of input conditioning unit and petri net control unit which is composed of the token control unit and firing unit. The firing unit determines the firing condition of the transfer signal on the basis of the token status of token control unit. By the proposed scheme, we can easily develop and implement the sequential system controller of automated warehousing system, automated transportation system, elevator system, and so on, as it is possible to modify control specification by changing simply the content of incidence matrix ROM and to expand easily functional capacity as the result of modular design.design.

• Proceedings of the Korea Institute of Convergence Signal Processing
• /
• 2000.12a
• /
• pp.205-208
• /
• 2000

In this paper, we'll show that an improved PDFF controller is obtained by substituting a feedforward compensator in the existing PDFF system with a dual-input describing function, and the controller has the ability of adjusting the bandwidth of a system as well as the phase margin simultaneously. The effectiveness of the proposed controller is confirmed by applying to the DC-motor position control system. As the results of simulation, we know that it is possible to design a controller by which the bandwidth of the closed system and its phase margin are easily adjusted.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.52 no.7
• /
• pp.119-124
• /
• 2015

Speech enhancement has been required as a preprocessor for a noise robust speech recognition system. Codebook-based Speech Enhancement (CBSE) is highly robust in nonstationary noise environments compared with conventional noise estimation algorithms. However, its performance is severely degraded for the codevector combinations that have lower correlation with the input signal since CBSE depends on the trained codebook information. To overcome this problem, only the reliable codevector combinations are selected to be used to remove the codevector combinations that have lower correlation with input signal. The proposed method produces the improved performance compared to the conventional CBSE in terms of Log-Spectral Distortion (LSD) and Perceptual Evaluation of Speech Quality (PESQ).

• Proceedings of the Korea Institute of Convergence Signal Processing
• /
• 2000.08a
• /
• pp.121-124
• /
• 2000

적응 학습 알고리즘으로 가중치를 변화시키는 단층 신경망의 출력부에 Ramp 임계 함수를 적용하여 입력이 zero-mean Gaussian random vector인 경우 가중치의 stationary point를 구하고, 적응 학습 알고리즘을 유도한다.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.41 no.5
• /
• pp.49-56
• /
• 2004

A low power consuming voice signal processing and control module was designed using a small $\mu$-controller for use in a totally implantable middle ear system. The module was designed that it can control the implanted system as well as process the fitting algorithm of input sound signal. In ordinary operation mode, the $\mu$-controller processes the applied sound signal for compensating the hearing loss of the patients. When the control signal is applied from the IR receiving module, the $\mu$-controller interrupts the signal processing and executes the order of the control signals such as power on/off, volume up/down. The designed module was implemented and verified the performance of the system through several experiments.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.12 no.3
• /
• pp.419-426
• /
• 2017

This paper describes an acoustic echo canceller with double-talk using BSS(: Blind Signal Separation) based on the geometric concept. The acoustic echo canceller may be deteriorated or diverged during the double-talk period. So we use the blind signal separation to detect the double talking by separating the near-end speech signal from the mixed microphone signal. In the closed reverberation environment, the blind signal separation extracts the near-end signal from unknown signals with the transformation and rotation based on the geometric concept. By this method, the acoustic echo canceller operates irrespective of double-talking. We verified performances of the proposed acoustic echo canceller by computer simulations. The results show that the acoustic echo canceller with this algorithm detects the double-talk periods thoroughly, and operates stably in the normal state without diverging of coefficients after ending the double-talking.

• Journal of the Korean Institute of Telematics and Electronics D
• /
• v.36D no.9
• /
• pp.6-12
• /
• 1999

It is difficult to realize adaptive main carrier rejection circuit in feedforward-type LPA(Linear Power Amplifier) because the gain and nonlinear characteristics of power amplifier are changed according to operating frequency, voltage, temperature. Usually, pilot tones are used for adaptive LPA operation. but in this paper, the relative phase, which in obtained through I&Q demodulator using input signals as LO signals and main-path & sub-path signals as RF signals, and the magnitude of main-path & sub-path signals are compared, so main carrier rejection is obtained. The proposed method rejects main carriers by 28.34 ~ 34.66dB (@Po=36.2 ~ 28.2 dBm/tone) with two tones at 877MHz, 882MHz and also rejects main carriers by 31.3dB despite changing condition of operating voltage.