Browsing by Author "Lee, Ming-Tang"

Now showing 1 - 7 of 7

An embedded audio-visual tracking and speech purification system on a dual-core processor platform
(2010-11-01)
Hu, Jwu-Sheng
;
Lee, Ming-Tang
;
Yang, Chia-Hsing
;
電控工程研究所
;
Institute of Electrical and Control Engineering
Design of an embedded audio-visual tracking and speech purification system is described in this paper. The system is able to perform human face tracking, voice activity detection, sound source direction estimation, and speech enhancement in real-time. Estimating the sound source directions helps to initialize the human face tracking module when the target changes the direction. The implementation architecture is based on an embedded dual-core processor, Texas Instruments DM6446 platform (Davinci), which contains an ARM core and a DSP core. For speech signal processing, an eight-channel digital microphone array is developed and the associated pre-processing and interfacing features are designed using the Altera Cyclone II FPGA. All the experiments are conducted in a real environment and the experimental results show that this system can execute all the audition and vision functions in real-time. (C) 2010 Elsevier B.V. All rights reserved.
Multi-channel post-filtering based on spatial coherence measure
(2014-12-01)
Hu, Jwu-Sheng
;
Lee, Ming-Tang
;
交大名義發表
;
電機工程學系
;
National Chiao Tung University
A multi-channel post-filtering algorithm using the proposed spatial coherence measure is derived. The spatial coherence measure evaluates the similarity between the measured signal fields using power spectral density matrices. In the proposed post-filter, the assumption of homogeneous sound fields is relaxed. Besides, multi-rank signal models can be easily adopted. Under this measure, the bias term due to the similarity of the desired signal field and the noise field is further investigated and a solution based on bias compensation is proposed. It can be shown that the compensated solution is equivalent to the optimal Wiener filter if the bias or the noise power spectral density matrix is perfectly measured. Simulations with incoherent, diffuse, and coherent noise fields and a local scattered desired source were conducted to evaluate the algorithms. The results demonstrate the superiority of the proposed bias compensated post-filter across different types of noise fields with a more accurate signal model. (c) 2014 Elsevier B.V. All rights reserved.
Norm-Constrained Capon Beamforming Using Multirank Signal Models With Kalman Filter Implementation
(2014-09-01)
Hu, Jwu-Sheng
;
Lee, Ming-Tang
;
交大名義發表
;
電機工程學系
;
National Chiao Tung University
The performance of the Capon beamforming is known to degrade dramatically in the presence of model mismatches, especially when the desired signal is present in the training data. To improve the robustness, the diagonal loading technique was introduced. However, the major drawback is the selection of the diagonal loading level, which is related to the unknown signal powers. Another way to alleviate performance degradation is to use a more accurate signal model of the array response. In this paper, a norm-constrained Capon beamforming using multirank signal models is proposed. Based on the pseudo-observation method, the quadratic constraints can be easily constructed. The problem is solved by a nonlinear Kalman filter which can be implemented on-line. The simulation results show that the design of the norm-constraint value is less sensitive to the signal powers, small angle mismatches, and number of sensors with a standard linear array. Further, it is shown that the use of a multirank signal model and Kalman filter technique result in less self-cancellation and performance degradation than that of the rank-1 signal model and the estimation of sample matrix.
Robust Adaptive Beamformer for Speech Enhancement Using the Second-Order Extended H-infinity Filter
(2013-01-01)
Hu, Jwu-Sheng
;
Lee, Ming-Tang
;
Yang, Chia-Hsin
;
應用數學系
;
電機工程學系
This paper presents a novel approach to implement the robust minimum variance distortionless response (MVDR) beamformer. The robust MVDR beamformer is based on the optimization of worst-case performance and provides an excellent robustness against an arbitrary but norm-bounded desired signal steering vector mismatch. For real-time consideration, the beamformer was formulated into state-space observer form and the second-order extended (SOE) Kalman filter was derived. However, the SOE Kalman filter assumes an accurate system dynamic and statistics of the noise signals. These assumptions limit the performance under uncertainties. This paper develops the SOE H-infinity filter for the implementation of the robust MVDR beamformer. The estimation criterion in the SOE H-infinity filter design is to minimize the worst possible effects of the disturbance signals on the signal estimation errors without a prior knowledge of the disturbance signals statistics. Experimental results demonstrate the performance of the proposed algorithm in a noisy and reverberant environment and show its superiority of the robustness against mismatches over the robust MVDR beamformer based on the SOE Kalman filter.
Simultaneous Localization of a Mobile Robot and Multiple Sound Sources Using a Microphone Array
(2011)
Hu, Jwu-Sheng
;
Chan, Chen-Yu
;
Wang, Cheng-Kang
;
Lee, Ming-Tang
;
Kuo, Ching-Yi
Sound source localization is an important function in robot audition. Most existing works perform sound source localization using static microphone arrays. This work proposes a framework that simultaneously localizes the mobile robot and multiple sound sources using a microphone array on the robot. First, an eigenstructure-based generalized cross-correlation method for estimating time delays between microphones under multi-source environments is described. Using the estimated time delays, a method to compute the far-field source directions as well as the speed of sound is proposed. In addition, the correctness of the sound speed estimate is utilized to eliminate spurious sources, which greatly enhances the robustness of sound source detection. The arrival angles of the detected sound sources are used as observations in a bearing-only simultaneous localization and mapping procedure. As the source signals are not persistent and there is no identification of the signal content, data association is unknown and it is solved using the FastSLAM algorithm. The experimental results demonstrate the effectiveness of the proposed method. (C) Koninklijke Brill NV, Leiden and The Robotics Society of Japan, 2011
Sound Source Tracking and Speech Enhancement by Microphone Array on an Embedded Dual-core Processor Platform
(2008)
Hu, Jwu-Sheng
;
Huang, Chi-Yang
;
Yang, Chia-Hsing
;
Wang, Chen-Kang
;
Lee, Ming-Tang
This paper proposes and implements a robot audition module which can detect voice activity, estimate sound source direction and enhance speech signal in a noisy environment These algorithms are implemented on an embedded dual-core processor platform which includes a ARM subsystem and DSP subsystem. An eight-channel digital microphone array is also implemented for sound data acquisition. All the experiments were conducted in a real environment and the experimental results show that this system can execute all the audition functions in real-time.
Wake-Up-Word Detection for Robots Using Spatial Eigenspace Consistency and Resonant Curve Similarity
(2011-01-01)
Hu, Jwu-Sheng
;
Lee, Ming-Tang
;
Wang, Ting-Chao
;
交大名義發表
;
National Chiao Tung University
In this paper, we propose a method to detect the wake-up-word (WUW) using microphone array for human-robot interaction. The consistency of the spatial eigenspaces formed by the speech source at different frequencies and the resonant curve similarity of the WUW are used as the features for detection. These features are processed and detected separately and the result is determined by cascading individual outcome using Bayes risk detector. This proposed method can keep a high recognition rate under very low signal-to-noise ratio (SNR) conditions. In addition, this method can estimate the direction of arrivals of the sound source, and the proposed architecture is easy to expand by adding detectors with other features in the cascaded manner to further improve the recognition rate.