at last compared two hardware realizing project : embedded project and dsp project . base the need of this system, i choose the dsp project . the dsp project adopted tms320vc5409-80 as main chip, match with ad converter tlc2558, power chip tps73hd318, to be the whole hardware circuit . at the same time realized the interface between speech recognition system and safety testing equipment . because there was some differences between voltage of dsp system and mega128 system, so i discussed how to convert voltage and get the speech recognition sub-system which can suited for original system . because the system get limit speech samples, and in order to ensure the stability of the system, so i add speech identification chip rsc364 as a sub system to ensure the stability of the system 同时实现该语音识别系统与原电器安全性测试系统接口,由于dsp系统与mega128系统存在电平转换的问题,就又将电平转换的方法进行了讨论,得到与原系统匹配的语音识别系统,由于本系统的识别算法获得的语音样本有限,为了确保本语音识别系统的稳定可靠,在语音识别系统中最后加入语音识别专用芯片rsc364作为辅助的语音识别子系统来保证系统的可靠性。