摘要
针对在PC机上实现的语音识别系统具有运行费用高、功耗大和体积大的问题,本文提出了以TMS320VC5509低功耗芯片为核心的语音识别系统方案,设计了系统的硬件结构,研究语音信号的特征提取和编码,建立前向-后向隐马尔可夫(HMM)模型,实现孤立词的语音识别。该系统具有运行费用低、功耗小和占用空间小的特点。
In view of the speech recognition system which realizes on PC machine has the problems of the operating cost is high, the power loss and the volume are big. This paper presents a speech recognition system based on the low power loss chip TMS320VC5509, designes the system＇s hardware architecture, researches the extracting of acoustic features and speech coding, sets up the Up-Back Hidden Markov Model（HMM） to realize speech recognition for small vocabulary. This system has advantages of operating cost is low and the power loss and the volume are small.
出处
《微计算机信息》
北大核心
2008年第23期197-199,共3页
Control & Automation
基金
"基于智能Agent技术的住宅无线网络应用研究"辽宁省教育厅项目(20060700)
