Linux论文：基于嵌入式的特定人语音识别智能轮椅设计.docVIP

Linux论文：基于嵌入式的特定人语音识别智能轮椅设计 【中文摘要】针对人们特别是残疾人愈来愈需要智能轮椅来改善他们的生活质量和生活自由度的需求。本文设计了一种物美价廉的智能轮椅来满足人们的需要。本文以S3C2440作为主控芯片,在Linux操作系统下,通过对电动轮椅增加由凌阳(SPCE061A)单片机、稳压电源电路、驱动电路组成的语音控制系统,实现用声音对电动轮椅的控制。该系统的软件采用隐马尔可夫模型(HMM)的特定语音辨识,可以有效的解决特定人的语音识别,拓展现有电动轮椅的使用人群的范围。其中,研究内容主要是语音识别部分,Linux系统下控制部分,以及基于QT/Embedded的图形用户界面。语音识别如果通过硬件来实现不仅比较困难,而且制作成本较高识别效果不好；选择通过软件来实现语音识别,这不仅降低了成本而且减少了元件缩小了整机的体积。 【英文摘要】people with disabilities more and more need for intelligent wheelchairs to improve their quality of life and freedom of life needs. This paper presents a smart affordable wheelchair to meet the needs of them.In the Linux operating system, S3C2440 as the master chip, voice control system is composed by increasing sunplus(SPCE061A) microcontroller, power supply circuit, drive circuit, to achieve using voice control of electric wheelchair. The system抯 software which uses Hidden Markov Model (HMM) speech recognition-specific.could effectively solve specific human speech recognition, and expand the use of existing electric wheelchair range crowd. Among them, the study is mainly included part of speech recognition, Linux system control section, and based on QT/Embedded graphical user interface. It is more difficult to achieve speech recognition by the hardware, meanwhile production costs is higher, recognition results are not satisfactory. Choose to implement voice recognition though software, this not only reduces the cost and component, the volume of the whole machine. 【关键词】Linux QT/Embedded HMM 语音识别 电动轮椅 【英文关键词】Linux S3C2440 QT / Embedded HMM Speech-Recognition SPCE061A electric-wheelchair 【目录】基于嵌入式的特定人语音识别智能轮椅设计 摘要 4-5 ABSTRACT 5 目录 6-7 第一章 绪论 7-10 1.1 设计背景 7 1.2 研究概况和发展趋势 7-8 1.3 语音识别存在的问题 8 1.4 本论文的主要内容和研究步骤 8-10 第二章 智能轮椅语音识别的理论基础及原理 10-20 2.1 语音学基础知识 10-12 2.2 语音识别的前端处理 12-17 2.3 语音特征提取 17 2.4 模型训练与模型匹配 17-20 第三章 嵌入式智能轮椅语音识别系统的硬件设计及实现 20-31 3.1 系统硬件电路的构成 20-21 3.2 主控模块S3C2440 21-22 3.3 电源模块 22-23 3.4 语音模块 23-24 3.5 存储模块 24-27 3.6 通信模块 27-29 3.7 电机模块 29-3