合肥工业大学硕士学位论文基于非特定人车载电器语音集成控制系统的设计与实现姓名：郭恒飞申请学位级别：硕士专业：信号与信息处理指导教师：王琼2011-04基于非特定人车载电器语音集成控制系统的设计与实现 基于非特定人车载电器语音集成控制系统的设计与实现 摘 要 摘 要 越来越多的电器加入到车载电器行列中，在改善汽车性能的同时，也增加了汽车驾驶操作的复杂度，甚至给行车过程带来了不安全隐患。其中，最主要的危险来自于驾驶人员在操作车载电器时注意力的转移， 造成交通事故的发生。新一代车载电器的研究面临着驾驶安全的问题，而采用语音控制车载电器是解决这一问题的有效方法。随着新一代语音专用系统芯片 Unispeech 的诞生和语音识别算法的优化，语音识别的识别率有了很大的提高，语音识别技术在汽车上的应用得以实现。采用语音控制可以在驾驶员不分散注意力的情况下，实现车载电器的控制，更可以进行快速操作，并且可以缓解驾驶员手动操作的压力，进而提高驾驶安全性。 本文提出了一种基于 HMM 算法非特定人车载电器语音集成控制方案。其中语音识别单元的处理核心采用 Infineon 公司推出的具有 DSP 和单片机双核的Unispeech-SDA80D51 语音处理专用芯片，语音执行单元的处理核心采用多块MCU 芯片， 并以整车车载电器为控制对象， 成功地实现对车载音响、 汽车空调、汽车雨刷、非安全性车灯和车窗的语音集成控制，并且针对车载环境内噪声干扰和电磁干扰进行了抗干扰设计，最终组成了车载语音控制系统，并制作了样机系统。 本文在样机系统制作完成后，测试样机系统的语音识别率和操作准确率，其中语音识别率达到了 92.71%， 操作准确率达到了 96%以上， 达到了预期目标。 关键词：关键词：车载电器 非特定人语音识别 语音控制 Unispeech-SDA80D51 HMM Design and Implementation of Automotive Electronic Equipment Control System based on Speaker-Independent Speech Recognition ABSTRACT More and more electrical equipments which could improve the performance of the automotive have been added to the automotive.However,it could not only makes drivering more harder but also brings more hidden dangers. One of the most dangers is the drivers attention away from the road when the drivers try to operate these electrical equipments,and many traffic accidents could happen because of this. New generation of automotive electrical equipments are facing the driving safety problems,and there is an effective way to solve this problem by using voice control. The speech recognition rate has been greatly improved as the new generation of voice-dedicated SoC Unispeech birth and optimization of speech recognition algorithms. And it has been achieved in many automotive applications. The drivers could oprate the automotive electrical equipment fast and easy without divert attention based on voice control.It makes drive more easy and safety. This paper presents a automotive electronic equipment control system which based on speaker-independent HMM algorithm speech recognition. The speech recognition unit based on Unispeech-SDA80D51,a speech recognition processing SOC witch has DSP and microcontroller dual-core inside. Voice processing execution unit based on chips of MCU.The control objects are automotive electrical equipments. Successfully achieved the automotive audio,car air conditioning, windshield wipers, non-safety lights and windows integrateds speech control. In addition，add anti-jamming design to system against noise and electromagnetic interference inside of automotive，leading to the formation of the automotive voice control system, and produced a model system. After the model system has be completed, we had tested the system performance and accuracy of operation on it, and the voice recognition rate reached 92.71%, the accuracy of operation rate reached 96%. It has proven that shis system achieved the expected goals. Keywords ： Automotive Electronic Equipment ； Speaker-Independent Speech Recognition；Speech Control System；Unispeech-SDA80D51；HMM 插图清单插图清单 图 2.1 帧移和帧长示意图 .7 图 2.2 三角滤波器之间的关系.7 图 2.3 HMM 组成示意图 .10 图 2.4 孤立词语音识别系统框图 .14 图 3.1 系统硬件总体设计 .16 图 3.2 Unispeech 语音识别基本框图 .17 图 3.3 Unispeech 开发板 IM220CN.24 图 3.4 样机系统中的语音执行单元电路板 .25 图 3.5 SL1102C1 型车载音响实物图 .25 图 3.6 SL1102C1 控制面板 KEY1P 接口电路 .26 图 3.7 SL1102C1 控制面板接口.26 图 3.8 POW 按键的模拟电路.26 图 3.9 车载音响语音执行单元电路 .26 图 3.10 SL06XXA1 型车载空调控制器实物图 .27 图 3.11 EC35AH 数字编码开关输出波形 .27 图 3.12 车载空调控制器语音执行单元主控电路.28 图 3.13 雨刷控制杆和雨刷电机实物图.29 图 3.14 永磁式雨刷马达电路.29 图 3.15 车载雨刷语音执行单元主控电路.30 图 3.16 车灯语音执行单元主控电路 .30 图 3.17 车灯电路与控制电路示意图 .31 图 3.18 车窗电路和控制电路实例 .32 图 3.19 汽车车窗语音执行单元主控电路.