• 대한임베디드공학회논문지
• /
• 제15권5호
• /
• pp.227-234
• /
• 2020

In this paper, we propose wav-U-Net to improve speech enhancement in heavy noisy environments, and it has implemented three principal techniques. First, as input data, we use 128 modified Mel-scale filter banks which can reduce computational burden instead of 512 frequency bins. Mel-scale aims to mimic the non-linear human ear perception of sound by being more discriminative at lower frequencies and less discriminative at higher frequencies. Therefore, Mel-scale is the suitable feature considering both performance and computing power because our proposed network focuses on speech signals. Second, we add a simple ResNet as pre-processing that helps our proposed network make estimated speech signals clear and suppress high-frequency noises. Finally, the proposed U-Net model shows significant performance regardless of the kinds of noise. Especially, despite using a single channel, we confirmed that it can well deal with non-stationary noises whose frequency properties are dynamically changed, and it is possible to estimate speech signals from noisy speech signals even in extremely noisy environments where noises are much lauder than speech (less than SNR 0dB). The performance on our proposed wav-U-Net was improved by about 200% on SDR and 460% on NSDR compared to the conventional Jansson's wav-U-Net. Also, it was confirmed that the processing time of out wav-U-Net with 128 modified Mel-scale filter banks was about 2.7 times faster than the common wav-U-Net with 512 frequency bins as input values.

• 한국음향학회지
• /
• 제35권4호
• /
• pp.280-287
• /
• 2016

본 논문에서는 다채널 능동 소음 제어를 위한 적응 필터에 적용할 수 있는 정규화된 FxLMS(Filtered-x Least Mean Square) 알고리즘을 제안하였다. 단일 채널 능동 소음 제어를 위한 FxLMS 알고리즘의 경우는 기존의 NLMS(Normalized Least Mean Square) 알고리즘과 같은 방식으로 정규화할 수 있는 반면, 다채널 능동 소음 제어의 경우에는 단일 채널 방식의 정규화 알고리즘을 그대로 적용할 수 없다. 먼저, 최소 교란 원리에 근거한 일반화된 정규화 알고리즘을 이용하여, 역행렬 연산을 피하기 위하여 대각 성분만을 고려한 정규화 알고리즘을 제안하였다. 컴퓨터 모의 실험을 통하여 제안된 알고리즘을 정규화되지 않은 기존의 알고리즘들과 비교하였다. 제안된 알고리즘이 정규화되지 않은 기존의 알고리즘에 비하여 비정상 환경에서 우수한 성능을 가진다는 것을 보였다.

• 대한전자공학회논문지SP
• /
• 제49권1호
• /
• pp.104-111
• /
• 2012

본 논문에서는 향상된 최소값 제어 재귀 평균 기법 (improved minima controlled recursive averaging, IMCRA) 알고리즘의 잡음 전력 추정성능을 향상 시키기 위한 알고리즘을 제안한다. 기존의 IMCRA은 주파수 특성이 빠르게 변화하는 비정상적인 환경과 낮은 SNR을 갖는 상황에서 잡음 전력 추정에 직접적으로 영향을 미치는 음성 검출기의 성능이 강인하지 못한 단점이 있다. 본 연구에서는 강인한 음성 검출 성능을 위해서 기존 IMCRA의 음성 검출기에 전역 음성 부재 확률을 적용한 음성 향상 기법을 제안한다. 제안된 알고리즘의 성능 평가는 음성의 perceptual evaluation of speech quality (PESQ)와 composite measure를 통한 음질을 평가하였다. 실험 결과 다양한 잡음 환경 (car, white, babble)에서 전역 음성 부재 확률을 적용한 IMCRA의 음성 향상 기법이 향상된 결과를 보여주었다. 특히, 비정상잡음 환경인 babble 5dB에서 PESQ 0.026, composite measure 0.029의 향상된 음질을 나타내었다.

• 한국소음진동공학회논문집
• /
• 제24권6호
• /
• pp.437-443
• /
• 2014

Vibration testing is conducted for evaluate the fatigue resistance of responsible system over excitation situations and two kinds of vibration profiles, harmonic or random, are widely used in engineering fields. Harmonic excitation profile is adequate for the rotating machinery that is primarily exposed to the orderly excited force subjected for a rotating speed; Random profile is suitable for the non-stationary vibration input, that is a ground excitation for example. Recently, the sine on random(SOR) testing method was sometimes considered to represent the real excitation conditions since the measured response signals of a target system, expecially for moving mobility, shows usually a mixture of them. So, it is important to understand the accumulated fatigue damage over different excitation patterns, harmonic and/or random, to determine the efficient vibration profile of a target system. A uniaxial vibration testing with a notched simple beam was introduced to evaluate the fatigue damage for different excitation profiles and the best choice of vibration profile was concluded from those comparison of calculated fatigue damages.

• 한국소음진동공학회:학술대회논문집
• /
• 한국소음진동공학회 2014년도 춘계학술대회 논문집
• /
• pp.340-345
• /
• 2014

Vibration testing is conducted for evaluate the fatigue resistance of responsible system over excitation situations and two kinds of vibration profiles, harmonic or random, are widely used in engineering fields. Harmonic excitation profile is adequate for the rotating machinery that is primarily exposed to the orderly excited force subjected for a rotating speed; Random profile is suitable for the non-stationary vibration input, that is a ground excitation for example. Recently, the sine on random (SOR) testing method was sometimes considered to represent the real excitation conditions since the measured response signals of a target system, expecially for moving mobility, shows usually a mixture of them. So, it is important to understand the accumulated fatigue damage over different excitation patterns, harmonic and/or random, to determine the efficient vibration profile of a target system. A uniaxial vibration testing with a notched simple beam was introduced to evaluate the fatigue damage for different excitation profiles and the best choice of vibration profile was concluded from those comparison of calculated fatigue damages.

• 한국음향학회지
• /
• 제36권1호
• /
• pp.64-69
• /
• 2017

본 논문에서는 PTT(Push-To-Talk) 기반의 무선 통신을 위한 진폭 및 위상 복원 기반의 단일 채널 음성 향상 방식을 제안한다. 제안한 방식은 신호의 진폭만을 대상으로 음성 향상을 진행했던 기존의 방식들과 달리, 음성 신호의 진폭과 위상을 분리하여 각각 향상시켜 다시 결합함으로써 더욱 양질의 음성을 제공한다. 본 논문에서 제안하는 방식의 성능을 평가하기 위해 동적 잡음 환경에서의 단계별 비교 실험을 실시하였으며, 실험 결과를 통해 제안한 방식이 다양한 잡음 환경에서 양질의 음성을 제공하는 것을 확인할 수 있다.

• 한국멀티미디어학회논문지
• /
• 제17권9호
• /
• pp.1064-1069
• /
• 2014

In this paper, we propose a novel approach to improve the performance of a voice activity detection(VAD) which is based on the adaptive band-partitioning with the likelihood ratio(LR). The previous method based on the adaptive band-partitioning use the weights that are derived from the variance of the spectral. In our VAD algorithm, the weights are derived from LR, and then the weights are incorporated with the entropy. The proposed algorithm discriminates the voice activity by comparing the weighted entropy with the adaptive threshold. Experimental results show that the proposed algorithm yields better results compared to the conventional VAD algorithms. Especially, the proposed algorithm shows superior improvement in non-stationary noise environments.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2004년도 학술대회지
• /
• pp.890-895
• /
• 2004

This paper describes constant alarm rates process of millimeter wave radar that exits on non-stationary target detection schemes in the ground clutter conditions. The comparison of various CFAR processes such as CA(Cell-Average)-CFAR, GO(Greatest Of)/SO(Smallest Of)-CFAR and OS(Order Statistics)-CFAR performance are applied. Using matlab software, we show the performance and loss between detection probability and signal to noise ratio. When rang bins increase, this results show the OS-CFAR process performance is better than any others and satisfies the optimal detection probability without loss of detection in the homogeneous clutter.

• 한국자동차공학회논문집
• /
• 제20권2호
• /
• pp.85-89
• /
• 2012

This paper presents a study on the system identification of the in-situ output shaft torque measurement system using a non-contacting magneto-elastic torque transducer installed in a vehicle drivline. The frequency response (transfer) function (FRF) analysis is conducted to interpret the dynamic interaction between the output shaft torque and road side excitation due to the road roughness. In order to identify the frequency response function of vehicle driveline system, two power spectral density (PSD) functions of two random signals: the road roughness profile synthesized from the road roughness index equation and the stationary noise torque extracted from the original torque signal, are first estimated. System identification results show that the output torque signal can be affected by the dynamic characteristics of vehicle driveline systems, as well as the road roughness.

• International Journal of Aeronautical and Space Sciences
• /
• 제6권1호
• /
• pp.8-16
• /
• 2005

This paper describes target detection performance of millimeter wave radar that exits on non-stationary target detection schemes in the ground clutter conditions. The comparison of various CFAR process schemes such as CA(Cell-Average)-CFAR, GO(Greatest Of)/SO(Smallest Of)-CFAR, and OS(Order Statistics)-CFAR performance are applied. Using matlab software, we show the performance and loss between target detection probability and signal to noise ratio. This paper concludes the OS-CFAR process performance is better than any others and satisfies the optimal detection probability without loss of detection in the homogeneous clutter, When range bins increase.