Font Size: a A A

Microphone Array Speech Enhancement And System Implementation For Human-machine Interaction

Posted on:2019-05-22Degree:MasterType:Thesis
Country:ChinaCandidate:J W LiFull Text:PDF
GTID:2428330545483796Subject:Physical oceanography
Abstract/Summary:PDF Full Text Request
Microphone array technology is a technique which uses multiple microphones to form a spatial structure to collect and process target sound source signals.Compared with the traditional single-microphone system,the microphone array system can collect the spatial information of the working environment and can effectively improve the performance of the speech signal processing.Therefore,the microphone array technology has become one of the research frontiers at home and abroad,and been widely used in practical applications such as car audio systems,interactive smart speakers,medical hearing aids,security monitoring systems,and audio/video conference systems.With the continuous development of society,the quality of human life have also been continuously improved.In recent years,various smart home products have entered the daily life of the family.Through using smart home product by human-computer interaction,human life is becoming more intelligent and more convenient.With the continuous development of human life intelligence,people's performance requirements for the microphone array working system are also rapidly increasing.However,under practial environment,the performance of the microphone array is subject to serious degradation at the presence of noise interference and reverberation.In the common office environment under the background of human-machine speech interaction,this paper focuses on the problem caused by noise interference,reverberation and other factors.A new dereverberation speech enhancement algorithm is proposed.Then a software and hardware systems is implemented.The main tasks of this article are:1.The basic principles of several common sound source localization algorithms are introduced.The performance of various sound source localization algorithms is verified through experiments in different environments.2.Based on the basic framework of the generalized sidelobe canceller speech enhancement algorithm,a sidelobe-enhancement reverberation mitigation algorithm is proposed to achieve dereverberation and noise reduction.Finally,the performance of the proposed algorithm is verified through experiments.3.Based on the STM32 embedded processor,a human-machine interaction microphone array working system is designed by using the proposed speech enhancement algorithm,to realize real-time localization and speech enhancement output of the target sound source.Finally,the working performance of the system is verified through experiments.
Keywords/Search Tags:microphone array, sound source localization, speech enhancement, beamforming, sidelobe enhancing
PDF Full Text Request
Related items