Research on Speech Recognition of Sanitized Robot Based on Improved Speech Enhancement Algorithm

In this paper, a voice enhancement algorithm is designed for the characteristics of being easily disturbed in the process of voice interaction recognition of sanitizing robots, in which the design process of the improved algorithm given, the ideas and methods of speech enhancement algorithm and keyword interval denoising are offered to improve the success rate of keyword speech recognition in noise and the design and application of speech enhancement algorithm are finally completed. As algorithm does not need to learn a lot in advance, it saves costs; improves the computing speed and portability, especially the iterative processing of keyword inter-word noise, reducing the false recognition rate of key control instructions. Finally, compared with the traditional spectral subtraction method, the algorithm is tested, and the improved spectral subtraction method, the noise reduction enhancement effect of the proposed algorithm obviously exceeds, its performance is better than that of the traditional spectral subtraction method, and the SNR is improved by an average of 4.40dB compared with the traditional spectral subtraction method. Especially compared with the spectrogram, which can better deal with the inter-word noise while maintaining the spectral characteristics of keywords. The effect is super as the original design goals were met, which is confirmed that the algorithm has good application and scalability, and is of high practical value.