外文翻譯--基于語(yǔ)音識(shí)別的智能門(mén)控系統(tǒng)設(shè)計(jì)（英文）

1、2009 ISECS International Colloquium on Computing, Communication, Control, and Management 978-1-4244-4246-1/09/$25.00 ©2009 IEEE

2、 CCCM 2009 Design and Realization of an Intelligent Access Control System Based on Voice Recognition Bo Cui College of Information Hebei Polytechnic University Tangshan,

3、China E-mail: mikecui@heut.edu.cn Tongze Xue College of Information Science and Engineering Hebei University of Science and Technology Shijiazhuang, China E-mail:dianzixinxi@yeah.netAbstract—The intelligent access

4、system adopts the technology of the voice recognition that based on the SPCE061A single chip. The system hardware is made up of SPCE061A single chip, the power and gating circuit, the extended memorizer SPR4096, the

5、voice input and output circuit. The key technologies are the application of the SPCE061A single chip on voice recognition and the design of the door control circuit. The system software is made up of the voice trainin

6、g module, the voice recognition module, the voice data processing module, the voice-playing module and the code of input and output module. The system completes the functions of collecting the voice data, distilling

7、character, special voice recognition and voice playing in terms of initializing the system and the identification training. According to the voice recognition arithmetic theory, the pretreatment of voice signal, the

8、character distilling and pattern matching is analyzed. The results of the experiments indicate that the system capability is steady and the identification is effective. The system can be applied in house or small offi

9、ce safety protection. Keywords-access control; linear prediction; voice recognition; pattern match; character distil I. INTRODUCTION There are many identification technologies used in current intelligent guard syste

10、m. Relative to other techniques, the voice recognition technology is generally regarded as one of the convenient and safe recognition techniques; this technique is a kind of technique that makes use of the creature c

11、haracter of a human body to carry on identification. Because everybody's creature character is unique and stable in a certain period, they are different from others and difficult to be fabricated and imitated. So

12、the voice recognition technology can be made use of identification, which is safe, accurate and dependable. II. THE CLASSIFICATION OF VOICE RECOGNITION Because the purpose and function of the voice recognition are

13、different, the recognition is classified as talker recognition and voice recognition. And the talker recognition can be classified as two types, one is relevant to text and the other is irrelevant to text. The voice r

14、ecognition system that is relevant to text needs users to pronounce according to the stated contents, and then everybody's speech model is built up accurately. Because identification also need users to pronounce a

15、ccording to the stated contents, the effect is very good. The voice recognition system that is irrelevant to text doesn't rule pronunciation contents of the talkers; it is difficult to build up speech models. But

16、 customers make use of the system conveniently, and it is applied widely. From the usage, the system can be classified as talker recognition and talker confirmation. The former judges a voice that needs to be identifi

17、ed from several talkers. The latter judges that an identified voice comes from a certain talker whether or not. Its output only has two kinds of result, yes or not. The central processor of this system is the SPCE061

18、A single chip. The talker confirmation that is relevant to text is realized on the chip, and then homologous order and operation are carried out. The system is mainly made up of talker identification module, gating c

19、ircuit and door lock etc. In training, the voice of talkers gets into the voice signal collection circuit through a microphone, and then the collected voice signals are processed by the voice processing circuit, the

20、characteristic parameters of talkers are distilled and saved. At last the database of characteristic parameters of talkers is formed. In identifying, the voice that needs to be identified is matched with the informati

21、on in the database of characteristic parameters of talkers. Output circuits control the gating electrical machine, and lastly the door lock is controlled. Figure1. Frame of this system 229most alike reference templa

22、te is found out, the voice is the identification result. A. Voice signals pretreatment The noises seriously disturb the processing and identification of voice signals, so the noises must be disposed firstly. The input

23、 analog voice signals from microphones must be sampled and measured in order to obtain digital voice signals. Before converting voice signals into digital signals, it is necessary to filter and counter- disturb. In

24、filtering the signal part and noises beyond 1/2 sample frequency are filtered. The clean voice signals are obtained later, and then low- frequency disturbing is filtered through fore-aggravation technology, especiall

25、y the disturbing of 50Hz or 60Hz. The high-frequency voice signals are improved and they can remove DC floating, restrain random noises and improve the function of energy of clean voice. B. Characteristic distilling T

26、he system adopts the evaluation method that uses the contrast value between dispersion degree of different speakers and self dispersion degree of each speaker as characteristic parameters. The basic idea is: to distil

27、l group characteristic parameters from a voice segment of the same speaker, that is to say to map the segment on a dot of the multi-space. Different voices from the same speaker will produce different dots in the cha

28、racteristic space; the function of multi-variable probability density can describe the distribution. For different single pronunciation from the same speaker, these dots are relatively concentrated. But the pronuncia

29、tion distribution from different speakers is apart farther, the group characteristic parameters can describe the thumbprint of speakers effectively. According to this principle, for single parameter, the F ratio betw

30、een two kinds of distribution parameters can be used as effective measurement rule. The F ratio shows the contrast between dispersion degree of different speakers and self dispersion degree of each speaker. The F rat

31、io of one characteristic parameter is bigger, for this characteristic, the former is bigger than the latter averagely. Therefore the recognition system adopts a bigger F ratio and then the system capability is improve

32、d. C. Module match At present research on the method of module match based on various characteristic parameters is more and more embedded. Typical methods are: the vector measure arithmetic, the Gauss mixture module a

33、rithmetic, the dynamic time whole (DTM) arithmetic and the manual nerve net arithmetic. The above methods have both advantage and weakness. When the DTM arithmetic is applied in the identification of long voice, the

34、operations of module match are too great. But the arithmetic is simple and effective for short voice (the length of valid voice is subter-3 seconds) identification. So the method is especially applicable to the speak

35、er recognition system of short voice and text. The system in this paper adopts the DTW arithmetic. V. EXPERIMENT RESULTS The users must confirm the practicability and stability of this system. Then two important para

36、meters of the system capability are the error identification ratio and the rejection identification ratio. The former is the mistake made by the voice signals of non-users accepted by the system, the latter is the mi

37、stake made by the voice signals of users rejected by the system, they are relevant to match threshold value. The setting of the match threshold value is relevant to the function and application of the voice lock syste

38、m. For family users the error identification ratio and the rejection identification ratio of this system must be likely low, even is zero. Firstly a certain user's voice is recorded by a recorder (the tone, the

39、speed and the content are all same), then the record is compared with the real voice forty times within different distance. It is showed in Table 1. After forty experiments by recorder the time of passage of uncertai

40、n users is zero in the results. The experiment result is very ideal for family users. When speakers are apart from the microphone farther the identification ratio descends obviously. The reason is that the ratio of th

41、e energy of useful voice signals and noise signals descends gradually; distortion appears when characteristic distilling and lastly leads to inaccessibility requests in module match. By adjusting the threshold value

42、the problem can be solved. VI. CONCLUSIONS The intelligent lock system in this paper adopts the voice recognition technology, namely the voice control intelligent system operates the lock system and gives correspondin

43、g voice answers and suggestions according to the voice signals. Compare with other biology identification technologies, the voice identification technology has such advantages as non-loss, non-forgetting, convenience

44、 etc. And also it is accepted by many users, the cost of voice input equipments is low. The system can be spread and applied conveniently because the system doesn't involve the users' privacy. Experiment test

45、s show that the system is stable and effective. The innovation in the paper is to apply voice processing technology of the SPCE061A single chip in the lock system, develop the usage scope of single chips, enrich appl

46、ication field of voice recognition and provide a new method for the intelligent lock system. REFERENCES [1] Jiqing Han, Lei Zhang, Tieran Zheng. Voice Signals Processing[M]. Beijing: Tsinghua University Press, 2004 [

47、2] Tiecheng Yu. The Development State of the Voice Identification [J]. The Development Communication World. 2005, 2:56-59 [3] Shiqiang Zhao, Bingli Jin, Man Zhao. Application and Exploitation of the 16 Bits Single Ch

48、ip Processor SPCE061A [J]. International Electronic Elements. 2003, 5:37-39 [4] Zhiling Jiang. The System of Data Multiprogramming Collection and Transmission Using Lingyang Singlechip Computer [J]. Sichuan Universit