• 한국컴퓨터정보학회논문지
• /
• 제26권2호
• /
• pp.61-68
• /
• 2021

최근 악성코드와 정상 바이너리를 분류하기 위해 기계학습을 이용하는 기술이 다양하게 연구되고 있다. 효과적인 기계학습을 위해서는 악성코드와 정상 바이너리를 식별하기 위한 Feature를 잘 추출하는 것이 무엇보다 중요하다. 본 논문에서는 재귀적인 방법을 이용하여 기계학습에 활용하기 위한 Feature 추출 방법인 RFA(Recursive Feature Addition) 제안한다. 제안하는 방법은 기계학습의 성능을 극대화 하기 위해 개별 Feature를 대상으로 재귀적인 방법을 사용하여 최종 Feature Set을 선정한다. 세부적으로는 매 단계마다 개별 Feature 중 최고성능을 내는 Feature를 추출하여, 추출한 Feature를 결합하는 방법을 사용한다. 제안하는 방법을 활용하여 Decision tree, SVM, Random forest, KNN등의 기계학습 알고리즘에 적용한 결과 단계가 지속될수록 기계학습의 성능이 향상되는 것을 검증하였다.

• 한국음향학회지
• /
• 제27권5호
• /
• pp.256-261
• /
• 2008

본 논문에서는 기존의 최소값 제어 재귀 평균기법(minima controlled recursive averaging, MCRA) 알고리즘에 조건 사후 최대 확률 (maximun a posteriori, MAP)을 적용한 음성향상을 제안한다. 기존의 MCRA는 파워스펙트럼에 평균을 취하고 각 서브밴드에서 음성 신호 존재 확률로 조절하는 스무딩 매개변수를 사용한다. 본 논문에서 제안된 알고리즘은 현재 프레임에 들어온 신호가 이전 프레임에서의 음성의 존재와 부재에 대한 조건을 부여해 주어 음성 신호 존재확률을 수정하여 음성향상에 적용한다. 제안된 음성 향상은 ITU-T P.862 perceptual evaluation of speech quality (PESQ)와 주관적 음질평가를 이용하여 평가하였고 기존의 MCRA 방법보다 향상된 결과를 나타내었다.

• International Journal of Control, Automation, and Systems
• /
• 제4권6호
• /
• pp.736-747
• /
• 2006

Simultaneous Localization and Map Building(SLAM) is one of the fundamental problems in robot navigation. The Extended Kalman Filter(EKF), which is widely adopted in SLAM approaches, requires extensive computation. The conventional particle filter also needs intense computation to cover a high dimensional state space with particles. This paper proposes an efficient SLAM method based on the recursive unscented Kalman filtering in an environment including a large number of landmarks. The posterior probability distributions of the robot pose and the landmark locations are represented by their marginal Gaussian probability distributions. In particular, the posterior probability distribution of the robot pose is calculated recursively. Each landmark location is updated with the recursively updated robot pose. The proposed method reduces filtering dimensions and computational complexity significantly, and has produced very encouraging results for navigation experiments with noisy multiple simultaneous observations.

• 한국정보과학회:학술대회논문집
• /
• 한국정보과학회 2005년도 가을 학술발표논문집 Vol.32 No.2 (1)
• /
• pp.922-924
• /
• 2005

실시간 시스템의 개발 및 운영에 사용되는 실시간 운영체제는 여러 개의 태스크가 동시에 작업할 수 있는 멀티태스킹 환경과 각 태스크에 우선순위를 부여하여 가장 높은 우선순위의 태스크가 CPU 를 선점하는 스케줄링 방법, 그리고 태스크간 동기화 및 통신을 위한 메커니즘을 제공하고 있다. 또한 여러 태스크들에 의해 사용되는 공유자원을 관리하기 위해 세마포어(Semaphore)를 사용하여 태스크간 동기화를 제공한다. 그러나 하나의 태스크가 세마포어를 이용하여 공유자원에 대해 여러 번 접근 할 경우, 데드락(Deadlock)을 일으킬 소지가 많다. 본 논문에서는 실시간 운영체제인 $iRTOS^{TM}$ 에서 데드락을 방지하기 위한 재귀적 세마포어(Recursive Semaphore)를 설계 및 구현하였다.

• ETRI Journal
• /
• 제25권2호
• /
• pp.144-147
• /
• 2003

We propose a new serial concatenation scheme for space-time and recursive convolutional codes, in which a space-time code is used as the outer code and a single recursive convolutional code as the inner code. We discuss previously proposed serial concatenation schemes employing multiple inner codes and compare them with the new one. The proposed method and the previous one with joint decoding, both performing a combined decoding of the simultaneous output signals from multiple antennas, give a large performance gain over the separate decoding method. In decoding complexity, the new concatenation scheme has a lower complexity compared with the multiple encoding/joint decoding scheme due to the use of the single inner code. Simulation results for a communication system with two transmit and one receive antennas in a quasi-static Rayleigh fading channel show that the proposed scheme outperforms the previous schemes.

• 전자공학회논문지B
• /
• 제31B권7호
• /
• pp.33-41
• /
• 1994

A single channel adaptive noise canceling (ANC) technique with a recursive time delay estimator (RTDE) is presented for removing effects of additive noise on the speech signal. While the conventional method makes a reference signal for the adaptive filter using the pitch estimated on a frame basis from the input speech, the proposed method makes the reference signal using the delay estimated recursively on a sample-by-sample basis. As the RTDEs, the recursion formulae of autocorrelation function (ACF) and average magnitude difference function (AMDF) are derived. The normalized least mean square (NLMS) and recursive least square (RLS) algorithms are applied for adaptation of filter coefficients. Experimental results with noisy speech demonstrate that the proposed method improves the perceived speech quality as well as the signal-to-noise ratio and cepstral distance when compared with the conventional method.

• 대한전기학회논문지:전기기기및에너지변환시스템부문B
• /
• 제48권3호
• /
• pp.139-143
• /
• 1999

This paper is on realization of the speed-sensorless vector control of an induction motor using the RLS(Recursive Least Square) algorithm. The speed estimator is including the RLS algorithm and a rotor flux observer. The RLS algorithm has speed and rotor time constant as parameter vectors and rotor flux observer is designed to have robustness to stator resistance variation and through the IP(Integral and Proportional) speed controller stable performance is obtained for estimating rotor speed. Finally the total algorithm are realized in induction motor drive system and its effectiveness is verified.

• 대한음성학회:학술대회논문집
• /
• 대한음성학회 2002년도 11월 학술대회지
• /
• pp.179-182
• /
• 2002

The development of noise robust speech processing systems is becoming increasingly important as speech technology is currently widely applied in real world applications. Recently, to resolve such a noise problem, adaptive noise canceller(ANC) is frequently used, which is based upon adaptive filters. The adaptive recursive filters perform better than adaptive non-recursive filters due to the added poles, but the stability may be severely threatened. But these problems of adaptive recursive filters was solved by ACHARF algorithm. This paper presents a method which combines speaker verification system with ANC(Adaptive Noise Canceller) using the ACHARF algorithm. In the front-end stage, ANC is adopted to suppress the additive noise imposed on the speech signal. The results show that the performance of speaker verification system becomes better than before.

• 제어로봇시스템학회논문지
• /
• 제6권12호
• /
• pp.1053-1060
• /
• 2000

In this paper, a discrete-time variable structure control method using recursively defined switching function and a decoupled variable structure disturbance compensator is used to achieve high performance circular motion control of a CNC machining center. The discrete-time variable structure control with the decoupled disturbance compensator method developed in this paper uses a recursive switching function defined as the sum of the current tracking error vector and the previous value of the switching function multiplied by a positive constant less than one. This recursive switching function provides much improved performance compared to the method that uses a switching function defined only as a linear combination of the current tracking error. Enhancements in tracking performance are demonstrated in the circular motion control using a CNC milling machine.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2011년도 제42회 하계학술대회
• /
• pp.1740-1741
• /
• 2011

This paper proposes a new linear recursive target state estimator for automotive collision warning system. The target motion is modeled in Cartesian coordinate system while the radar measurements such as range, line-of-sight angle and range rate are obtained in polar coordinate system. To solve the problem by nonlinear relation between these two coordinate system, a practical linear filter design scheme employing the predicted line-of-sight Cartesian coordinate system (PLCCS) is proposed. Especially, PLCCS can effectively incorporate range rate measurements into target tracking system. It is known that the utilization of range rate measurements enables the improvement of target tracking performance. Moreover, PLCCS based target tracking system is implemented by linear recursive filter structure and hence is more suitable scheme for the development of reliable collision warning system. The performance of the proposed method is demonstrated by computer simulations.