基于DSP的车载语音识别系统设计  

Design of a speech recognition system in cars based on DSP

作　　者：王磊[1] 郝润科[1] 赵龙[1] 邱银[1] 刘源[1] 

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海200093

出　　处：《信息技术》2013年第6期33-35,40,共4页Information Technology

基　　金：上海市研究生创新基金项目(JWCXSL1202)

摘　　要：为了提高在汽车噪声环境中语音识别系统的识别率,并达到降低系统成本的目的,在低成本定点DSP(TMS320VC5509A)芯片工业标准的基础上,使用两路无差输入信号进行噪声抵消,建立了一种基于孤立词,小词汇量,特定人识别的语音识别系统,实验表明,在类似汽车噪声环境中,该系统的识别率可满足实用要求,且成本较低。In order to improve the recognition rate of the speech recognition system in the noise environment of the motor vehicle, and achieve the purpose of reducing the cost of the system, this paper used two input signals to implement adaptive noise cancellation, and designed a particular person effectively speech recognition system with single and few words on the basis of the industry standard of the low cost and f＇Lxed-point DSP chip, TMS320VCJJ09A. The experiments showed that the recognition rate of the system in similar noise environment of the motor vehicle can achieve the practical requirements even with a lower cost.

关 键 词：TMS320VC5509A 语音识别 自适应噪声抵消 

分 类 号：TN911.7[电子电信—通信与信息系统]