论文总字数：34832字

摘 要

智能家用垃圾桶是家居智能化发展的产物，它对于室内垃圾分类工作有着举足轻重的作用，本文设计中的智能家用垃圾桶产品搭载了语音识别系统，使得在识别垃圾种类时更为便捷、准确和智能化。随着社会的不断发展和进步，垃圾分类逐渐实现普及化和智能化，各类的智能家用垃圾桶层出不穷，对于智能家用垃圾桶的研究也更加具有重要意义。

而根据课题背景研究，许多智能垃圾分类桶采取告诉用户如何投放垃圾，由用户进行主观投放的方式，这种方式使垃圾分类具有不稳定性，因此本次智能垃圾分类桶产品采用单独分类，单独投放的方式完成垃圾分类功能，具体为智能垃圾桶拥有四种盛放不同种类垃圾的分桶，在智能垃圾桶识别垃圾名称之后进行内部分类，然后转动垃圾桶盖至用户应当投放的垃圾分桶内，这部分功能主要靠本文中的语音识别系统设计实现。为此，本文进行了以下研究工作：

1. 研究传统的语音识别流程、声学模型HMM的建立以及语音识别算法原理，如DTW算法、VAD技术等。
2. 硬件平台选型结果为STM32F103系列单片机和LD3320语音识别芯片，并基于Proteus 8软件以及STM32开发板对语音识别系统进行硬件设计，完成主电路以及辅助电路的搭建。
3. 基于Keil uVision 5对软件系统进行代码编写与调试，舍弃传统的循环识别，而选用更为稳定的触发识别方式，研究ASR流程。
4. 通过串口助手软件使计算机与STM32开发板进行串口通讯，仿真其语音识别功能，并对语音识别率进行测试，根据结果研究变量，对语音识别率进行一定的优化调整。
5. 基于Proteus 8和Keil uVision 5软件联合仿真STM32驱动28系列步进电机，实现了电机根据语音识别结果进行定位旋转的功能。

本课题最终完成了包含47个垃圾名称的语音库建立，通过对语音音量、识别距离、语速、MIC增益、环境等多个因素进行综合考量，设定出最佳的语音识别方案，有效提高了室内环境下语音识别的成功率，基本实现了预期语音识别系统应实现的功能。

关键词：智能家用垃圾桶；语音识别系统；STM32；LD3320；电机驱动

Abstract

Intelligent household garbage can is the product of the development of household intelligence, which plays an important role in the classification of indoor garbage. The intelligent household garbage can product in this paper is equipped with a voice recognition system, which makes it more convenient, accurate and intelligent in the identification of garbage types. With the continuous development and progress of society, garbage classification has gradually achieved popularization and intelligence, and all kinds of intelligent household garbage cans emerge in endlessly, which is more important for the research of intelligent household garbage cans.

According to the background research of the project, many intelligent garbage bins take the way of telling users how to put in garbage and subjectively putting it in by users, which makes the garbage classification unstable. Therefore, this intelligent garbage bin product uses the way of separate classification and separate putting to complete the garbage classification function, specifically, the intelligent garbage bin has four kinds of different kinds of garbage The garbage bin is classified internally after the intelligent garbage bin identifies the garbage name, and then the garbage bin cover is rotated to the garbage bin that the user should put in. This part of function is mainly realized by the design of the speech recognition system in this paper. For this reason, this paper has carried out the following research work:

1.This paper studies the traditional speech recognition process, the establishment of acoustic model HMM and the principle of speech recognition algorithm, such as DTW algorithm, VAD technology and so on.

2.The result of hardware platform selection is STM32F103 series single chip and ld3320 speech recognition chip. Based on Proteus 8 software and STM32 development board, the hardware of speech recognition system is designed, and the main circuit and auxiliary circuit are built.

3.Based on keil uVision 5, the software system is coded and debugged. Instead of the traditional cycle recognition, the more stable trigger recognition method is used to study the ASR process.

4.Through the serial assistant software, the computer communicates with STM32 development board through serial port, simulates its speech recognition function, tests the speech recognition rate, and optimizes the speech recognition rate according to the result research variables.

5. Based on Proteus 8 and keil uVision 5 software, a 28 series stepping motor driven by STM32 is simulated, which realizes the function of motor positioning and rotation according to the result of speech recognition.

In this project, the speech database including 47 garbage names is finally established. Through the comprehensive consideration of voice volume, recognition distance, speech speed, MIC gain, environment and other factors, the best speech recognition scheme is set up, which effectively improves the success rate of speech recognition in indoor environment, and basically realizes the expected functions of speech recognition system.

Key Words：Intelligent household garbage can; speech recognition system; STM32; LD3320; motor drive

目录

第1章 绪论...................................................................................................................................1

1.1 课题背景及意义..................................................................................................................1

1.2 国内外研究现状..................................................................................................................1

1.2.1 国外研究现状.............................................................................................................1

1.2.2 国内研究现状.............................................................................................................2

1.3 预期目标..............................................................................................................................4

1.4 本设计主要工作..................................................................................................................4

1.5 全文组织结构图..................................................................................................................4

第2章 智能家用垃圾桶总体规划...............................................................................................6

2.1 功能需求分析......................................................................................................................6

2.2 总体方案..............................................................................................................................7

第3章 智能家用垃圾桶的语音识别系统...................................................................................8

3.1 语音识别系统结构与流程..................................................................................................8

3.1.1 语音识别系统结构.....................................................................................................8

3.1.2 语音识别流程.............................................................................................................8

3.2 语音识别系统声学模型及算法研究..................................................................................9

第4章 面向语音识别系统的硬件设计.....................................................................................10

4.1 硬件平台总体框架............................................................................................................10

4.2 硬件平台设备选型............................................................................................................10

请支付后下载全文，论文总字数：34832字