就目前三种主流的语音识别算法：动态时间规(DTW)、隐马尔科夫模型(HMM)和人工神经网络(ANN)。分析它们的原理、特点及实现过程，对 DTW 的语音识别进行实验，通过对比分析三种算法的特点，结合本文研究的实际情况，选择 DTW 作为研究的重点，提出利用遗传算法对其进行改进。

-The three mainstream speech recognition algorithms: Dynamic Time Regulations (DTW), hidden Markov model (HMM) and artificial neural network (ANN). Analysis of their principles, characteristics and implementation process, DTW speech recognition experiment, by comparing the characteristics of the three algorithms, combined with the actual situation of this study, select the DTW as the focus of the study, proposed the use of genetic algorithms to improve it.