语音识别电风扇控制系统设计

摘 要

本文论述了一个采用语音识别技术的电风扇控制系统，采用凌阳16位单片机微处理器作为整个系统的核心，且由语音输入输出电路以及电风扇驱动电路组成。本系统能识别特定人的语音命令自动完成对风扇启动、关闭以及档位切换的控制。在本语音识别控制系统中充分利用凌阳单片机内部的10位高精度AD资源采集特定人控制命令信号，同时也充分利用其内部10位高精度DA资源转换处理好的语音信号进行输出，这样不用外加AD、DA芯片，不仅节约了成本，而且也简化了整个系统的设计。在特定人发出语音命令后，单片机接收语音信号并与语音参考样本进行对比识别，其中识别对比采用的是DTW动态时间规整算法。识别成功后，单片机给相应I/O口发出电平驱动控制电风扇工作运转。设计中加入了液晶显示模块，让命令者能及时地知晓风扇的工作状态。

针对目前语音识别系统普遍存在的缺陷：识别正确率低，本文所设计系统的亮点是在一定程度上提高语音识别的正确率，能较准确听从特定人命令完成对电风扇启动关闭以及档位切换的控制。本文所设计的系统还有一个特色是实用性很强，通过适当的硬件电路改造，该语音识别系统还可以应用于其他家用电器的控制中，且本系统相对于市场上其他语音识别产品，其设计成本较低。

关键词：语音识别；SPCE061A；电风扇控制系统

Abstract

This paper mainly discusses a control system of electric fan which depends on the speech recognition. This control system applies the Sunplus 16 SCM microprocessor as its operating core and is made up of voice input and output circuit, as well as electric fan drive circuit. The speech recognition control system fully utilizes the 10 high-precision AD resources in the Sunplus SCM SPCE061A to collect the certain control command signal. At the same time, the voice signal which disposed by 10 high-precision DA resources are fully applied to output data. In this way, it is no need to add the AD and DA chips, so that more costs can be reduced and design of system can be simplified. When the given person launches the voice command, SCM begins to receive the voice signal and conduct the comparison identification with voice samples. The comparison identification applies DTW dynamic time algorithm. After the identification finishing, SCM issues the electrical level to the corresponding I/O, in order to control the operation of electronic fan. In the design process, the LCD Module is used, which allows the commander know the working condition of electronic fan in time.

In terms of the existing drawbacks of the current speech recognition—low accuracy rate of identification, the polishing point of this paper is that the accuracy rate of speech recognition system will be improved and the commander’s control of swi