Papers by Wataru Mitsuhashi
電子情報通信学会技術研究報告 Sip 信号処理 Ieice Technical Report, Mar 1, 2012
電子情報通信学会技術研究報告 Ie 画像工学, Jan 3, 2007
Signal Processing, Jun 1, 2009
The presence of strong acoustic feedback degrades the performance of active noise control (ANC) s... more The presence of strong acoustic feedback degrades the performance of active noise control (ANC) systems, and in the worst case the ANC system may become unstable. In this paper, we investigate additive-random-noise-based methods for online feedback path modeling and neutralization (FBPMN) in multichannel ANC systems. In existing methods, separate filters are used for each acoustic feedback path: one for online feedback path modeling (FBPM), and another for feedback path neutralization (FBPN). Furthermore, the existing method works well for predictable noise sources but its performance degrades for unpredictable sources. This paper attempts to solve these problems in multichannel ANC systems. The proposed method is a modification and extension of previous work for online FBPMN in single-channel ANC systems. In the proposed method, we combine the two tasks of FBPM and FBPN in one online FBPMN filter, and hence, the computational complexity of the proposed method is lower than existing methods. Computer simulations are carried out to demonstrate the effectiveness of the proposed method. It is shown that the proposed method achieves better performance than existing methods for both predictable and unpredictable noise sources. Furthermore, this improved performance is achieved at somewhat reduced computational complexity.
Sensors and Materials, 1998
電子情報通信学会技術研究報告 Ea 応用音響, May 22, 2008
電子情報通信学会技術研究報告 Ie 画像工学, Jan 5, 2009
電子情報通信学会技術研究報告 Sip 信号処理, Mar 1, 2012
電子情報通信学会技術研究報告 Ie 画像工学, Dec 31, 2007
Journal of Sound and Vibration, Nov 1, 2009
This paper concerns active noise control (ANC) of impulsive noise modeled using non-Gaussian stab... more This paper concerns active noise control (ANC) of impulsive noise modeled using non-Gaussian stable processes. The most famous filtered-x least mean square (FxLMS) algorithm for ANC systems is based on minimization of variance of error signal. For the impulse noise, the FxLMS ...
信号処理, Jul 1, 2010
ABSTRACT In this paper, we propose a new score function for subband decomposition ICA to improve ... more ABSTRACT In this paper, we propose a new score function for subband decomposition ICA to improve the algorithm performance. Subband decomposition ICA is known to having better separation performance than frequency-domain ICA. However, it is basically time-domain ICA, with much shorter filters, performed on each subband. In this paper, discussion is focused on an information-maximization based approach which performance depends on the distribution assumption of source signals. We propose the use of a generalized Cauchy distribution as a new distribution assumption to derive a new score function. With the proposed score-function, the performance of subband decomposition ICA algorithm is significantly improved in terms of both SIR and convergence speed.
2008 SICE Annual Conference, 2008
ABSTRACT
2008 2nd International Conference on Signal Processing and Communication Systems, 2008
ABSTRACT
This paper concerns the problem of uncorrelated periodic disturbance appearing at the error micro... more This paper concerns the problem of uncorrelated periodic disturbance appearing at the error microphone in singlechannel feedforward active noise control (ANC) systems. This disturbance, being uncorrelated with the primary noise, cannot be controlled by standard filtered-x least mean square (FxLMS) algorithm, and increases the residual noise. Furthermore, it significantly degrades the steady-state performance of ANC systems. Here we propose a new method which can simultaneously control both the correlated and uncorrelated noise signals. The proposed method comprises three adaptive filters; 1) the FxLMS algorithm-based ANC filter to cancel the primary noise, 2) a separate FxLMS algorithm-based ANC filter to cancel the uncorrelated disturbance, and 3) the LMS algorithm based supporting adaptive filter to generate appropriate signals for the two ANC filters. Computer simulations are carried out which demonstrate that the proposed method can effectively mitigate the correlated and uncorrelated primary disturbances, and gives significantly improved noise reduction performance. This improved performance is achieved at an increased computational complexity.
Spatial filter velocimetry (SFV) is often used in the fields of remote sensing and industrial ins... more Spatial filter velocimetry (SFV) is often used in the fields of remote sensing and industrial instrumentation for measuring the velocity of image motions by using optical computation. SFV is powerful at finding homogeneous translation in the image plane of optical devices. However, a SFV sensor requires a wide field of view for averaging image motion, thereby resulting in degradation in spatial resolution. In contrast, our SFV approach is based on a bank of Gabor reticles as spatial filters, where the reticle elements are regularly distributed on the image plane for determining the velocity field so that velocity components can be isolated from image motions. Applying the proposed SFV system to velocity estimation for a rigid target moving in 3D space enables us to visually track the target with a monocular camera system. A Kalman filter is also introduced to predict the target's position smoothly and improve the delay characteristics and reduce the error of localization. We empirically show that the SFV system in conjunction with a Kalman filter gives good results for detecting a target entering into the image plane and tracking it visually with good accuracy. 3D motion detection of a preceding vehicle has been among the subjects of major interests in intelligent transportation systems (ITS). The fundamental issue in extracting 3D mo- tion information in ITS is to determine the distance to the preceding vehicle in real-time operation. For the distance measurement, laser range-finders or RADAR systems are typically exploited, and stereoscopic camera systems are also adopted. In this paper, a velocity sensor for estimating 3D motion based on the spatial-filter velocimetry (SFV) in conjunction with a Kalman filter is proposed. The motion information is determined by using spatial Gabor reticles regularly dis- tributed in concentric configurations on the camera's image plane. Each reticle modulates the image intensities and the amplitude of the reticle's output fluctuates temporally ac- cording to the image motion, so that the velocity measure- ment can be reduced to frequency estimation. The motion parameters over the image plane, such as translation, rota- tion, and dilation, can be isolated by evaluating the frequency distribution over the bank of Gabor reticles. The system first detects a target entering into the field of view of the monocular camera, and then locates and fixates the target within the central area of the image plane. The size of the target image can be kept constant by detecting the dilating component of the image motion, thereby providing the information relevant to the target distance. A camera with three degree of freedom (pan, tilt, and zoom) was employed and its attitude was controlled accord- ing to the estimated motion parameters. A Kalman filter is adopted in the experiments to smoothly predict the target's position and improve the delay characteristics and reduce the error of localization. The proposed system is composed of only a commercially available PC without any specialized hardware. The system shows the ability of tracking a mov- ing planar target successfully at the operation time of about 30 frames per second with a slight positional error.
Proceedings of the 41st SICE Annual Conference. SICE 2002., 2002
ABSTRACT
IEEJ Transactions on Electronics Information and Systems
A new configuration of spatial reticles is described for applications in real-time object trackin... more A new configuration of spatial reticles is described for applications in real-time object tracking task. The motion of a planar object in a three-dimensional environment is decomposed into global motion parameters by measuring the temporal frequency of the reticles arranged on concentric circles in the image plane. Experiments were performed to maintain the size and the position of the object's image by controlling the attitude of a camera on the basis of the estimated motion parameters. The proposed system, composed of only a commercially available PC, exhibited good tracking performances with processing rate of 15 frames per second.
Uploads
Papers by Wataru Mitsuhashi