基于单片机的智能语音识别系统设计 本科生毕业设计论文 论文.doc

基于单片机的智能语音识别系统设计 （硬件部分） 系 别： 专 业 班： 姓 名： 学 号： 指导教师：  基于单片机的智能语音识别系统设计(硬件部分) The Design of Intelligent Speech Recognition System Based on Single-chip Computer (HardWare)摘 要 本文设计一个让机器通过识别和理解过程把语音信号转变为相应的文本或命令的高技术的语音识别系统。本语音识别系统以LD3320语音识别芯片为核心部件，主控MCU选用STC10L08XE。主控MCU通过控制LD3320内部寄存器以及SPI flash实现语音识别和对话。通过麦克风将声音信息输入LD3320进行频谱分析，分析后将提取到的语音特征和关键词语列表中的关键词语进行对比匹配，找出得分最高的关键词语作为识别结果输出给MCU，MCU针对不同的语音输入情况通过继电器对语音命令所对应的电器实现控制。同时也可以通过对寄存器中语音片段的调用，实现人机对话。 设计中，电源模块采用3.3V供电，主要控制及识别部分采用LM1117-3.3稳压芯片，语音播放及继电器部分采用7812为其提供稳定的电流电压。寄存器采用一片华邦SPI flash芯片W25Q40AVSNIG，大小为512Kbyte。系统声音接收模块采用的传感器为一小型麦克风——驻极体话筒，在它接收到声音信号后会产生微弱的电压信号并送给MCU。另外系统还采用单片机产生不同的频率信号驱动蜂鸣器来完成声音提示，此方案能完成声音提示功能，给人以提示的可懂性不高，但在一定程度上能满足要求，而且易于实现，成本也不高。 关键词： Abstract This paper designs a hi-tech speech recognition system which enables machines to transfer speech signals into corresponding texts or orders by recognizing and comprehending. The centerpiece of the speech recognition system is LD3320 voice recognition chip，its master MCU STC10L08XE. Master MCU achieve voice conversation by controlling the internal registers and SPI flash LD3320.The sound information is inputted into LD3320 by microphone to do spectrum analysis. After analyzing the voice characteristics extracted are compared and matched with the key words in the list of key words.Then the highest scores of key words found would be output to MCU as recognition results. MCU can control the corresponding electrical real of speech recognition for different voice input through the relays and can also achieve voice conversation through a call to voice clips in register. In the design, power module uses 3.3V.The main control and identification part adopt LM1117-3.3 voltage regulator chip,and 7812 is used to provide stable current and voltage for the part of voice broadcast and relay.Register uses chip SPI flash W25Q40AVSNIG which is 512Kbyte. The sensor used in the speech reception module of the design is microphone,namely electrit microphon