Blind DOA Estimation in a Reverberant Environment Based on Hybrid Initialized Multichannel Deep 2-D Convolutional NMF With Feedback Mechanism

Qiang Fu,Bo Jing,Pengju He

doi:10.1109/access.2019.2958955

Abstract

The accuracy performance of traditional direction of arrival (DOA) estimation algorithms is seriously affected by the reverberation. Considering the advantage of the sparse characteristic of speech signal in time-frequency (T-F) domain, this paper presents a new blind DOA estimation method based on integrated deep learning and convolutional non-negative matrix factorization (NMF). Firstly, mathematic models of microphone array and room impulse response are built. In addition, we extracted blindly initialization parameters of 2-D convolutional NMF using k-means clustering algorithm and singular value decomposition algorithm, which can be used to accurately estimate the main components of desired sound source in the reverberation environment of multi-path propagation. Moreover, the feedback mechanism is introduced into deep 2-D convolutional NMF and correlation coefficient between the signal decomposed by NMF and the signal to be decomposed is used to select the best separated signal for DOA estimation, which make the separation algorithm simpler and more efficient. Finally, test of orthogonality of projected subspaces (TOPS) algorithm is used to validate the DOA estimation capability of this algorithm. Compared with the unprocessed reverberation speech, the estimation error is reduced, which shows that the proposed algorithm can effectively improve the estimation accuracy of DOA estimation when the received signals are in a reverberant environment.

Highlights

Direction of arrival (DOA) is a fundamental theoretical problem in array signal processing (ASP), and is used widely in sonar, radar, mobile communication and electronic countermeasure [1]
The subspace-based algorithm is represented by MUSIC and ESPRIT, Schmidt developed the multiple signal classification (MUSIC) based on spatial spectral estimation in 1979, and Roy and Kailath developed the estimation of signal parameter via rotational invariance techniques (ESPRIT) in 1986
Blind single-channel speech dereverberation method based on N-CTF model and negative matrix factorization (NMF) is presents to enhance the quality of speech signals that have been recorded in an enclosed space [7]

Summary

INTRODUCTION

Direction of arrival (DOA) is a fundamental theoretical problem in array signal processing (ASP), and is used widely in sonar, radar, mobile communication and electronic countermeasure [1]. MUSIC algorithm is more accurate, more stable and higher angular resolution than ESPRIT algorithm These methods mentioned above are based on the premise that the measured signal is a pure signal, and can not apply to signals in strong reverberation environment directly. The problem of background noise has been well solved, but room reverberation still affects the performance of the algorithm to a large extent. Blind single-channel speech dereverberation method based on N-CTF model and NMF is presents to enhance the quality of speech signals that have been recorded in an enclosed space [7]. In view of the problems in the above research, this paper proposes a method based on deep 2D convolutional NMF to suppress reverberation in narrow enclosed space and achieve accurate DOA estimation.

PROBLEM DESCRIPTION

DEPTH EXTRACTION METHOD

TOPS ALGORITHM

TECHNIQUE PROCESS OF PROPOSED ALGORITHM

VIII. CONCLUSION