毕业论文-语音识别的研究.docVIP

语音识别的研究 摘要： ?与机器进行语音交流，让机器明白你说什么，这是人们长期以来梦寐以求的事情。语音识别技术就是让机器通过识别和理解把语音信号转变为相应的文本或命令的技术。而我们对语音识别系统的研究主要由三部分组成：主控系统、语音录入播放部分、执行部分；由主控系统发出命令，接着进行语音播报，执行部分执行主控命令。 本系统采用SUNPLUS公司生产的一款集16位MCU, A/D，AGC, 2KRAM， 32K Flash等在一块芯片上，且具有DSP功能的SPCE061A单片机。由它来完成语音识别及对直流电机的控制。本语音识别系统除了识别准确度高、系统体积小、耗电省、价格低、性能可靠的特点以外，还具有较快的运算速度，可实现实时的语音识别。并且本系统的语音提示功能为人机交流提供了一个良好界面。以上的特点充分体现了嵌入式语音识别的优势。本系统在进行语音训练的基础上可以实现特定人的语音识别和人机对话，如在此基础上进一步改进，会实现非特定人的语音识别，将更为完善。 关键字：语音识别，主控系统，SPCE061A单片机，人机交流 Research of Speech Recognition Author :Yang ling Pang ming Tutor : Pei dong Wang quanzhou ( Year:2006th falls due Major: Electronic information project Class:In 2002 level electronic first class Number:200272020138 200272020126) Abstract： Human being earnestly long for the communication and understanding with machines through speech. Speech recognition means the high technology which lets machines understand the speech on basis of the text or command converted from speech signal. The research field of speech recognition system which this paper focuses on is made up of three main parts including host-control system, speech record play and implementation. Firstly host-control system sends out command, and the secondary play record part broadcast this speech, then the host-control command executed by the implementation part. Host-control system adopts SPCE061A single-chip made by SUMPLUS. SPCE061A implements the speech recognition and controlling of servo motor. For instance, when sent out the command “BROTHER”, the system prompts to be in speech recognition. While sent out the command “Turn left”, the servo motor implements through the turning of specific angles. The system will go on with the repeated recognition in that way. Although at current time the system has not achieved the unrestricted communication between human being and machines, its advantages are simple operation, convenient control and expanding. KEY WORDS: speech recognition, host-control system, servo motor, spee