Font Size: a A A

Voice Recognition Capabilities In The Wireless Master Disk Test System Design And Implementation

Posted on:2008-12-05Degree:MasterType:Thesis
Country:ChinaCandidate:P C ShaoFull Text:PDF
GTID:2208360212975412Subject:Signal and Information Processing
Abstract/Summary:PDF Full Text Request
The speech recognition technology is the important developing direction of thecomputer technology, as an interdisciplinary field, it is valued in theoretic research andpromising in application. Speech recognition system has gained wide applications inrecent years due to rapid technology developments of digital signal processing.On the basis of studying the speech recognition technology overall and deeply, thedissertation firstly introduces the development of speech recognition, the dissertationelucidates the background and significance of the research and the difficulties we facedand introduces the principle of speech recognition, expatiate on the classification ofspeech recognition system and the method of recognition. Finally the dissertation takesthe method of DTW to design the word-isolated, speaker-dependent and small-scalespeech recognition system.Secondly, based on the composition of speech recognition model, the dissertationintroduces the steps in model-matching, which include pretreatment, endpoint detection,feature selection, and elucidates the principles of speech signal processing and patternrecognition that involve in parts of system.The dissertation discusses the possibility thatthe design of speech recognition on PC. With help of the theory, the dissertation takespretreatment of speech signal, and detects the endpoint of speech signal with the methodof double- threshold. It takes MFCC as the feature of speech signal by the way of LPand uses the method of DTW to compare model feature sequences and speech featuresequence. The program is designed in Visual C++ and Matlab. At last, the dissertationtakes out perspective in developing the method of DTW, and the new results promise tofurther improve application of speech recognition technology.The dissertation develops the platform of the test system and designs the programof communication controlling.Last part of the dissertation summarizes what has been accomplished and pointsout further research directions.
Keywords/Search Tags:speech recognition, endpoint detection, feature selection, dynamic time warping, serial port communication
PDF Full Text Request
Related items