• 한국산학기술학회:학술대회논문집
• /
• 한국산학기술학회 2008년도 추계학술발표논문집
• /
• pp.164-166
• /
• 2008

소음은 현대 사회에서 쉽게 접하게 되는 환경 오염원이다. 능동소음제어(Active Noise Control)는 발생된 소음을 제거하기 위해 구현이 간단한 LMS 알고리즘을 많이 사용하고 있다. 그러나 LMS 알고리즘은 수렴 속도와 소음신호의 변화속도에 따라 발산의 위험을 가지고 있다. 본 논문에서는 이러한 LMS의 문제점을 보완하기 위해 경험 모드 분석법을 이용한 feedback FXLMS(Filtered-X Least Mean Square) 알고리즘을 제안하였다. 소음제거 시스템의 출력단에서 검출된 잔차소음을 경험 모드 분석법(Empirical Mode Decomposition)을 이용하여 IMF 신호들로 분해하고, 분해된 각 신호를 FXLMS 알고리즘을 이용하여 수렴시킨 후, 결과들을 다시 결합하여 소음 제거에 이용하였다. 각각의 IMF 신호를 FXLMS 알고리즘으로 수렴시킬 때 수렴속도에 변화를 주어 소음제거의 효율성을 높였다. 제안한 알고리즘을 Matlab을 이용하여 시뮬레이션하였고 기존의 FXLMS알고리즘보다 향상된 수렴속도 및 안정성을 가짐을 입증하였다.

• 전력전자학회논문지
• /
• 제4권6호
• /
• pp.589-600
• /
• 1999

사무실에서 사용되는 소용량 무정전진원 장치(UPS : Uninterruptible Power Supply)는 대부분의 소음이 냉각팬을 통하여 발생되고 있는데, 이와 같이 장치의 여러 방향으로 방사되는 소음의 능동적 소음제어를 위해서는 FXLMS(Filtered-x LMS) 알고리즘에 기초를 둔 다중채널 방식의 FXLMS(multiple-channel FXLMS) 알고리즘의 사용이 적당하다. 본 논문에서는 사무실에서 주로 사용되는 소용량 UPS의 소음 발생 특성을 분석하고 다중채널 방식 MRMO(Multiple-Reference/Multiple-Output) FXLMS 알고리즘을 적용하여 방사상으로 전파되는 UPS의 소음을 저감시키기 위한 능동 소음 제어를 수행하고 이의 소음저감 특성을 고찰한다.

• 한국정밀공학회지
• /
• 제20권11호
• /
• pp.100-107
• /
• 2003

This paper represents a disturbance compensation control for attenuating disturbance responses. In the consideration of the requirements on the model accuracy in the model based compensator designs, an experimental feed forward compensator design based on adaptive estimation by Filtered-x least mean square (FXLMS) algorithm is proposed. The convergence properties of the FXLMS algorithm are discussed and its conditions for the asymptotic convergence are derived theoretically. The effectiveness of the proposed method and the theoretical proof are verified by computer simulation.

• 한국소음진동공학회:학술대회논문집
• /
• 한국소음진동공학회 2005년도 춘계학술대회논문집
• /
• pp.488-493
• /
• 2005

In gun stabilization systems, the torque comes from the unbalance mass of gun and the base acceleration is an important source of disturbance which degrades stabilization performance. Fatigue of gear train is another important factor affecting structural safety problems. In this paper, a feedback control gain is designed by optimal control weighting to difference between motor and gun velocity, and a feedforward controller using FXLMS algorithm is adopted to investigate those problems. Experimental results show that the feedforward compensator based on FXLMS can reduce the disturbance effects. The directional convergence property according to initial conditions of the FXLMS is also shown through experiments.

• 대한전기학회논문지:시스템및제어부문D
• /
• 제54권1호
• /
• pp.40-47
• /
• 2005

This paper concerns on Filtered-x least mean square (FXLMS) algorithm for adaptive estimation of feedforward control parameters. The conditions for convergence in ensemble mean of the FXLMS algorithm are derived and the directional convergence properties are discussed from a new geometric vector analysis. The convergence and its directionality are verified along with some computer simulations.

• 한국소음진동공학회:학술대회논문집
• /
• 한국소음진동공학회 2008년도 추계학술대회논문집
• /
• pp.245-246
• /
• 2008

The Correlation FXLMS (Co-FXLMS) algorithm was developed to improve the control performance. The Co-FXLMS algorithm is realized by using an estimate of the cross correlation between the adaptation error and the filtered input signal to control the step size. In this paper, the performance of the Modified Co-FXLMS is presented in comparison with that of the CO-FXLMS algorithm. Simulation results show that active noise control using Modified Co-FXLMS is effective to control performance of algorithm and prevent divergence.

• 한국소음진동공학회:학술대회논문집
• /
• 한국소음진동공학회 2009년도 추계학술대회 논문집
• /
• pp.240-241
• /
• 2009

The method of the reduction of the duct noise can be classified by the method of passive control and the method of active control. However, the passive control method has a demerit to reduce the effect of noise reduction at low frequency (below 500Hz) range and to be limited by a space. Whereas, the active control method can overcome the demerit of passive control method. The algorithm of active control is mostly used the Least-Mean-Square (LMS) algorithm because the LMS algorithm can easily obtain the complex transfer function in real-time. Especially, When the Filtered-X LMS (FXLMS) algorithm is applied to an ANC system.

• 전기학회논문지
• /
• 제61권10호
• /
• pp.1489-1495
• /
• 2012

In this study, the active noise control experiment has been performed using induction motor noises. While the noises were measured, a induction motor was operated in different speed. For the simulation of ANC(Active Noise Control), test-bed is composed a multi-channel ANC system was constructed. In order to compare the control performance, we performed noise reduction simulations of ANC by Co-FXLMS algorithm and FXLMS algorithm. Through the simulation results, we confirmed that convergence performance and noise decrease effect of the proposed Co-FXLMS algorithm have been improved from existing FXLMS algorithm.

• 한국소음진동공학회:학술대회논문집
• /
• 한국소음진동공학회 2004년도 춘계학술대회논문집
• /
• pp.780-783
• /
• 2004

Recently Intake noise has been extensively studied to reduce the engine noise. In order to diminish intake noise several resonators were added to the intake system. However this can cause a reduction of engine output power and an increase of fuel consumption. In this study, active noise control simulation of the Filtered-x LMS algorithm is applied real instrumentation intake noise data under rapid acceleration because intake noise is more excessively increased under the such a harsh condition. But the FXLMS algorithm has poor control performance when the system is disturbed. Thus modified FXLMS algorithm using L-point running average filter is developed to improve the control performance under the rapid acceleration and disturbance. The noise reduction quantity of modified Filtered-x LMS algorithm is more than original one in two cases. In the case of control for real instrumentation intake noise data, maximum residual noise of modified FXLMS algorithm is 2.5 times less than applied the FXLMS and also in the case of disturbed, the modified FXLMS algorithm shows excellent control performance but FXLMS algorithm cat not control.

• 전자공학회논문지T
• /
• 제35T권2호
• /
• pp.48-55
• /
• 1998

광대역 소음의 능동소음제어(Active Noise Control, ANC)에 사용되는 Filtered-X Least Mean Aquares(FXLMS) 알고리즘의 주 결점은 제어 알고리즘의 입력으로 사용되는 필터링된 기준신호가 상관되어 입력신호의 자기상관행렬이 큰 고유치비를 갖게 됨으로써 수렴속도가 느려지는 것이다. 이 상관은 그 원인이 기준센서 신호의 자기상관 때문이거나, 기준신호가 오차경로로 필터링되어서 나타난다. Recursive Least Squares(RLS)와 같은 빠른 수렴속도를 갖는 적응 알고리즘이 존재하지만, 이들 알고리즘은 실 계의 구현을 위해 많은 계산을 필요로 하거나, 각 샘플마다 제어기의 최적화가 필요한데, 이는 매우 심각한 제한조건이다. 본 논문에서는 빠른 수렴의 속도를 갖으면서 FXLMS와 근사한 연산량을 갖는 제어알고리즘을 제안한다. 이 알고리즘은 수렴속도를 느리게 하는 상관성을 줄이기 위해 이산 Walsh 변환을 사용하는 변환영역 FXLMS이다. 제안한 알고리즘을 광대역 능동소음제어 시스템에 적용, 모의시험한 결과 FXLMS의 문제점인 고유치비가 약 1/40으로 줄어드는 성능 개선 효과를 확인할 수 있었다.