基于语音识别的机械臂控制系统设计  

Design of the Manipulator Control System Based on Speech Recognition

作　　者：白梦飞 金衎哲 符朝兴[1] BAI Mengfei;JIN Kanzhe;FU Chaoxing(College of Mechanical and Electrical Engineering,Qingdao University,Qingdao 266071,China)

机构地区：[1]青岛大学机电工程学院,山东青岛266071

出　　处：《青岛大学学报(工程技术版)》2024年第4期44-51,共8页Journal of Qingdao University(Engineering & Technology Edition)

基　　金：山东省自然科学基金资助项目(ZR2020QE183)。

摘　　要：为了方便工作人员控制机械臂,设计了一种通过语音命令机械臂完成动作的控制系统。通过自动语音识别模块采集语音,对语音信号进行降噪处理,提高语音信号质量,利用语音识别模型对语音信号进行识别分析和关键词匹配,自动生成指导机械臂运动的数据。机械臂控制模块分析运动数据,并转化为控制指令,通过控制算法完成机械臂的运动动作。实验结果表明,谱减法降噪使语音识别准确率提高了14.77%,验证了降噪算法的有效性,满足系统对语音识别的技术要求。语音控制机械臂实验中,机械臂能够完成预期动作,验证了系统的可行性。To facilitate the control of the manipulator by staff, a control system has been proposed that enables the execution of actions via speech commands. An automatic speech recognition (ASR) module is utilized to collect speech, and the speech signal undergoes denoising to enhance its quality. A speech recognition model is employed to recognize, analyze, and match keywords within the speech signal, thereby automatically generating data to guide the movement of the manipulator. The manipulator control module processes the motion data and converts it into control instructions, executing the manipulator’s movements through control algorithms. Experimental results demonstrate that spectral subtraction denoising improves the accuracy of speech recognition by 14.77%, verifying the effectiveness of the denoising algorithm and fulfilling the technical requirements of the system for speech recognition. In the experiment of the speech-controlled manipulator, the manipulator can successfully complete the expected actions, confirming the feasibility of the system.

关 键 词：自动语音识别 谱减法 关键词匹配 机械臂控制 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]