改进的智能机器人语音识别方法(英文)

Improved speech recognition method for intelligent robot

作为一种人机信息交互技术,语音识别技术得到了广泛的应用。介绍了基于凌阳十六位单片机SPCE061A的语音识别系统,并且采用了以传统的线性预测倒谱系数(LPCC)与分形维数相结合的混合参数作为特征参数的语音识别方法。LPCC方法是体现说话人特定的声道共振特性的线性预测方法,而分形维数则可以定量的描述语音气流中的非线性混沌特征。实验结果表明,基于LPCC与分形维数混合参数的语音识别方法要比单一的LPCC参数语音识别方法识别效果好。

As a communication technology between man machine interactive technology, speech recognition is widely used. In this paper, we introduces the speech recognition system based on the 16 bit SPCE061A serial single chip and propose a speech recognition approach with mixed parameter, which combines the traditional Linear Predictive Cepstral Coefficients （LPCC） and fractal feature as the feature parameter. The LPCC method is linear procedure based on the assumption that speaker features have properties caused by the vocal tract resonances. Fractal dimension is used to quantitatively describe the chaos nonlinearity in speech air flow. The experimental results show that mixed feature parameter of LPCC and fractal dimensinn is better than single LPCC feature parameter in recognition rate.