HMM算法语音识别控制机械手的系统设计  

System Design of Control Manipulator Based on HMM Voice Recognition

作　　者：张起浩 将少国 赵鹏 Zhang Qihao;Jiang Shaoguo;Zhao Peng(Department of Electronic Science,Nankai University Binhai College,Tianjin 300270,China)

机构地区：[1]南开大学滨海学院电子科学系

出　　处：《单片机与嵌入式系统应用》2019年第10期82-85,共4页Microcontrollers & Embedded Systems

基　　金：天津市国家级大学生创新创业训练计划项目资助——特种环境下车载可拆卸机械臂(201813663001)

摘　　要：提出了基于HMM语音识别控制机械手的设计,在简述HMM算法模型原理的基础上,介绍了以S3C6410作为微处理器的硬件系统设计,以及嵌入式Linux移植的步骤。最后将LD3320语音识别模块和基于HMM算法语音识别分别在安静和嘈杂环境下对机械手进行控制。实验结果表明,基于HMM语音识别正确率高于LD3320语音识别模块,HMM算法在非特定人连续语音有着良好的识别性和抗噪性。The design of control manipulator based on HMM voice recognition is proposed in the paper.First of all,the model and principle of HMM algorithm are introduced.Then,the hardware system design of S3C6410 as microprocessor is introduced,and the porting steps of embedded Linux are listed.Finally,the LD3320 voice recognition module and the voice recognition module based on HMM algorithm are used to control the manipulator in the quiet and noisy environment respectively.The experiment results show that the HMM voice recognition has higher accuracy than the LD3320 voice recognition module,and the HMM algorithm can offer great recognition and anti-noise performance in speaker-independent continuous speech.

关 键 词：HMM LD3320 麦克风阵列 ARM LINUX移植 

分 类 号：TP32[自动化与计算机技术—计算机系统结构]