针对语音识别系统在噪声环境下识别率下降的问题,提出改进的感知非均匀谱压缩特征提取算法,它能在噪声环境下对语音信号进行有效的压缩,并且使训练与识别环境更加匹配,最终达到提高识别率的目的.在智能轮椅平台上进行语音识别验证,结果表明,该算法能有效地提高语音识别系统的鲁棒性,保证语音识别系统在噪声环境下的识别率. Aiming at the problem that the recognition rate of speech recognition system decreases in noisy environment, an improved perceptual non-uniform spectral compression feature extraction algorithm is proposed, which can effectively compress the speech signal under noisy environment and make the training and recognition environment more suitable. Finally, And achieve the goal of improving the recognition rate.The result of verification on the intelligent wheelchair platform shows that the algorithm can effectively improve the robustness of the speech recognition system and ensure the recognition rate of the speech recognition system in noisy environments.