Design, Implementation And Performance Improvement Of Human-computer Voice Interaction Hardware System Based On WM897

After years of rapid development,computers have penetrated into all aspects of human life with an irreversible trend,and has gradually become an indispensable part of the development of human society with its successful applications in embedded systems,edge computing,artificial intelligence algorithms and other fields.As the bridge between human and computer,human-computer interaction system not only determines the way of human-computer interaction,but also decides the efficiency of human-computer information exchange.Human-computer interaction is an important research interests in the application of computer technology.It mainly refers to human-computer information information exchange between people and computers through graphics,contact,voice,posture and other interactive methods to complete specific collaborative tasks.At present,humancomputer interaction mainly involves video and voice interaction.Human-computer video interaction mainly studies the interactive processing technology of image video,body posture and other aspects,which is a research hot spot.And human-computer voice interaction focuses on the research of human-computer voice communication technology,and there are comparatively few special studies in this area.However,voice interaction has been increasingly used because of its contactless interaction characteristics and the advantages of close to human natural language interaction.Because speech signals are easily disturbed by environmental noise,which can reduce the accuracy and reliability of human-computer interaction,the research of human-computer speech interaction system with strong anti-interference ability and strong robustness is an important part of humancomputer speech interaction research at present.This can not only improve the efficiency and accuracy of human-computer voice interaction system,but also expand the application scope of human-computer voice interaction system,especially in the special working environment where noise cannot be eliminated.Therefore,the research of this topic has important engineering significance.At first,this paper designs a human-computer voice interaction front-end hardware acquisition platform based on STM32 and WM8978,and gradually completes the overall logic planning of hardware,the design of printed circuit board,and the soldering and debugging of components.Besides,this paper has successfully completed the program design of the initialization of the front-end hardware audio acquisition system and the thread-safe audio acquisition algorithm,so that the platform can provide multiple formats of lossless audio data sets for the back-end audio data processing system.This hardware acquisition platform is also an important aspect of the research,which can assist in verifying the effectiveness of the theoretical algorithm proposed by this paper when dealing with real sampled signals.Secondly,on the basis of the existing speech processing model,combined with the main existing problems represented by frequency shift and validity of features in the speech signal processing,this paper proposes a back-end voice signal processing algorithm,which includes a frequency shift prediction method based on spectrum expectation and a feature evaluation scheme based on time-frequency information entropy.In order to verify the validity of the speech signal processing algorithm proposed in this paper,the algorithm has carried out experimental verification on the lossless audio data collected by the hardware acquisition platform under the noise environment and other experimental noise data,and has effectively and accurately displayed the characteristics of the noise,providing a perspective for noise removal.Besides,the experimental data verification also successfully verified the proposed methods can achieve the frequency shift prediction and suppression of spectrum features and the judgement of the validity of features.The hardware and software based research proposed by this paper improves the sampling accuracy and recognition accuracy of the front-end sampling system of humancomputer voice interaction in engineering,and optimizes and improves the noise recognition and removal of the back-end voice signal processing intelligent algorithm in theory.The research of this paper can provide reference experience for the wide application of human-computer voice interaction in noisy environment and other applications.