Dezvoltarea de algoritmi robusti pentru imbunatatirea semnalului vocal cu mai multe microfoane pentru sistemele de comunicatie din automobil

Robust Multi-Microphone Speech Enhancement Algorithms for Automobile Communications Systems

PN-II-RU-TE-2014-4-1880

Last update: 1/10/2017

Project Summary: For audio systems integrated into a car there are many applications that have not been sufficiently exploited and prospects of development of research. An important application for hands-free communication systems or voice commands recognition is to separate speech signal and reduce the noise signal to obtain an enhanced undistorted and intelligible voice signal. Another future application for a car audio system is amplifying the voice signal coming from the driver or front passenger to be heard over the rear passengers. Voice driver amplification system, however, can get into audio feedback with the rear speakers sound system, therefore it is necessary to cancel the acoustic feedback. This project aims at the development of robust systems to improve speech signals processed by audio communication systems of the automobile to increase comfort and reduce the risk of accidents. There will be studied multi-microphone noise reduction techniques, for which both spectral and spatial characteristics of the signal sources can be used. Also there will be studied adaptive algorithms for noise reduction using multi-microphones and for acoustic feedback cancellation. The developed methods will be real-time implemented and tested on hardware platforms with DSP and FPGA in order to be able to integrate it into the car audio system.

Research team:

1. prof.dr.ing. Radu Mihnea Udrea – director proiect

2. prof.dr.ing. Constantin Paleologu – cercetator cu experienta

3. s.l.dr.ing. Cristian Lucian Stanciu – tanar cercetator (doctor)

4. drd.ing. Adrian Patrascu – student doctorand

5. drd.ing. Razvan Florentin Trifan – student doctorand

Project objectives

Objectives		Associated tasks
1	Developing beamforming techniques for spatial focusing on the signal source	1.1. Study of time-delay estimation between the microphone signals for the acoustic source localization
		1.2. Study of the fixed beamforming technniques to enhance signal coming from a specific direction
		1.3. Designing far-field fixed broadband beamformers with an arbitrary desired spatial directivity pattern for a given arbitrary microphone array configuration
		1.4. Extension of the design of beamformers for a near-field broadband beamformer operating at one specific distance
		1.5. Create a simulation framework using Matlab and a multi-microphone hardware acquisition platform to test the proposed methods
		1.6. Research dissemination with national and international conferences papers and at least one ISI Journal paper
2	Developing multi-microphone signal enhancement techniques based on Generalized Singular Value Decomposition (GSVD)	2.1. Extension of the single-microphone speech enhancement techniques for multi-microphone optimal filtering techniques
		2.2. Techniques for reducing the computational complexity of the GSVD-based optimal filtering technique
		2.3. Incorporate the GSVD-based optimal filtering technique in a Generalised Sidelobe Canceller (GSC) structure with an adaptive noise cancellation (ANC) postprocessing stage
		2.4. Analyze of the performance of the GSVD-based optimal filtering technique for several simulated acoustic environments and for real-life recordings using the simulation framework
		2.5. Research dissemination with national and international conferences papers
3	Integration of multi-microphone signal processing and adaptive channelwise spectral subtraction	3.1. Study of the methods of integrating spectral subtraction (SS) and delay-and-sum beamforming (DS)
		3.2. Study of the differences of applying SS before or after DS
		3.3. Adaptive adjustment of the subtraction parameter for channelwise spectral subtraction
		3.4. Simulation and performance analysis of the proposed SS and DS multi-microphone solutions
		3.5. Real-time implementation of multi-microphone signal enhancement and adaptive channelwise spectral subtraction
		3.6. Research dissemination with national and international conferences papers and at least one ISI Journal paper
4	Analysis and development of adaptive algorithms for canceling acoustic feedback of automobile audio systems	4.1. The study of classical noise cancellation based on continuous adaptation reaction loop (CAF)
		4.2. The estimation of the transfer function of the acoustic feedback inside the automobile
		4.3. Evaluation of adaptive algorithms when imposing constraints on adaptive filter coefficients CAF system
		4.4. Simulation and real-time implementation of the system for cancelation of the acoustic feedback of car audio amplification systems
		4.5. Dissemination of research results through participation in national and international conferences and by filing a patent application for the proposed system

Dissemination results:

Book chapter:

· C. Anghel, C. Stanciu, and C. Paleologu, “Efficient FPGA implementation of a CTC turbo decoder for WiMAX/ LTE mobile systems,” in Field-Programmable Gate Array. George Dekoulis (Editor), InTech, Rijeka, Croatia, 2017.

Journal papers:

· J. Benesty, C. Paleologu, and S. Ciochina, “On the identification of bilinear forms with the Wiener filter,” IEEE Signal Processing Lett., vol. 24, pp. 653-657, May 2017.

· C. Anghel, C. Stanciu, and C. Paleologu, “LTE turbo decoding parallel architecture with single interleaver implemented on FPGA,” Circuits, Systems & Signal Processing, 21 pages, http://link.springer.com/article/10.1007/s00034-016-0362-z.

· C. Paleologu, S. Ciochina, J. Benesty, and S. L. Grant, “An overview on optimized NLMS algorithms for acoustic echo cancellation,” EURASIP Journal on Advances in Signal Processing, vol. 2015:97, pp. 1-19, Dec. 2015. http://www.asp.eurasipjournals.com/content/2015/1/97

Conference papers:

· C. Paleologu, J. Benesty, and S. Ciochina, “An NLMS algorithm for the identification of bilinear forms,” in Proc. European Signal Processing Conference (EUSIPCO), 2017, pp. 2689-2693, Kos, Greece.

· S. Ciochina, C. Paleologu, and J. Benesty, “Analysis of an LMS algorithm for bilinear forms,” in Proc. IEEE International Conference on Digital Signal Processing (DSP), 2017 (5 pages), London, UK.

· L. Dogariu, S. Ciochina, C. Paleologu, J. Benesty, and P. Piantanida, “An adaptive solution for nonlinear system identification,” in Proc. IEEE International Symposium on Signals, Circuits and Systems (ISSCS), 2017 (4 pages), Iasi, Romania.

· R. A. Dobre, V. A. Nita, S. Ciochina, and C. Paleologu, “Improved convergence model of the affine projection algorithm for system identification,” in Proc. IEEE International Symposium on Signals, Circuits and Systems (ISSCS), 2017 (4 pages), Iasi, Romania.

· R. A. Dobre, C. Paleologu, S. Ciochina, C. Negrescu, and D. Stanomir, “Investigation on the performances of APA in forensic noise reduction,” in Proc. IEEE International Conference on Speech Technology and Human-Computer Dialogue (SpeD), 2017 (6 pages), Bucharest, Romania.

· I. Albu, C. Anghel, and C. Paleologu, “Adaptive filtering in acoustic echo cancellation systems – A practical overview,” in Proc. IEEE International Conference on Electronics, Computers and Artificial Intelligence (ECAI), 2017 (6 pages), Targoviste, Romania.

· C. Elisei-Iliescu and C. Paleologu, “Recursive least-squares algorithms for echo cancellation - An overview and open issues,” in Proc. International Conference on Networks (ICN), 2017, pp. 87-91, Venice, Italy. *Paper Award*

· C. Elisei-Iliescu, C. Stanciu, C. Paleologu, J. Benesty, C. Anghel, and S. Ciochina, “Robust variable-regularized RLS algorithms,” in Proc. IEEE HSCMA, 2017, pp. 171-175, San Franciso, USA.

· C. Stanciu, C. Anghel, C. Paleologu, S. Ciochina, and J. Benesty, “FPGA implementation of an optimized NLMS algorithm,” in Proc. IEEE International Symposium on Electronics and Telecommunications (ISETC), 2016 (4 pages), Timisoara, Romania.

· R. A. Dobre, C. Elisei-Iliescu, C. Paleologu, C. Negrescu, and D. Stanomir, “Robust audio forensic software for recovering speech signals drowned in loud music,” in Proc. IEEE SIITME, 2016 (4 pages), Oradea, Romania.

· S. Ciochina, C. Paleologu, J. Benesty, S. L. Grant, and A. Anghel, “A family of optimized LMS-based algorithms for system identification,” in Proc. European Signal Processing Conference (EUSIPCO), 2016, pp. 1803-1807, Budapest, Hungary.

· C. Elisei-Iliescu, C. Paleologu, and R. Tamas, “On the performance of variable forgetting factor recursive least-squares algorithms,” in Proc. SPIE ATOM-N, 2016 (6 pages), Constanta, Romania.

· C. Paleologu, J. Benesty, C. Stanciu, C. Anghel, and M. Stenta, “Robust regularization of the recursive least-squares algorithm,” in Proc. IEEE International Conference on Electronics, Computers and Artificial Intelligence (ECAI), 2016 (4 pages), Ploiesti, Romania. *Invited Paper*

· C. Stanciu, C. Anghel, C. Paleologu, S. Ciochina, and J. Benesty, “On the numerical properties of an optimized NLMS algorithm,” in Proc. IEEE International Conference COMMUNICATIONS (COMM), 2016 (4 pages), Bucharest, Romania.

· S. Ciochina, C. Paleologu, J. Benesty, and C. Anghel, “An optimized affine projection algorithm for acoustic echo cancellation,” in Proc. IEEE Int. Conf. Speech Technology and Human-Computer Dialogue (SpeD), Bucharest, Romania, Oct. 14-17th, 2015, pp. 159-164. http://www.sped2015.ro/

· S. Ciochina, C. Paleologu, J. Benesty, and S. L. Grant, “An optimized proportionate adaptive algorithm for sparse system identification,” in Proc. IEEE Asilomar Conf. Signals, Circuits, and Computers, Pacific Grove, CA, USA, Nov. 8-11th, 2015, pp. 1546-1550. http://www.asilomarsscconf.org/