摘要
文章以机载语音控制的实时性、准确性、方便性的实际应用要求为出发点,深入讨论既适合于机载语音人机交互特性,又符合机载语音识别技术特点的语音指令的设计思想。提出机载语音控制的命令语言应该是建立在在现有汉语陆空通话语言语法及驾驶舱显控系统按键功能提示信息词汇基础之上,通过修改补充,进而形成一套有功能完整、富有层次的机载语音控制指令系统。文章同时提出,通过恰当选择控制指令的词汇读音和发音力度,能够为语音识别器提供更充分的识别信息量,提高声音信号的信噪比,有助于其在机载噪声和变异的条件下发挥出最佳的识别性能。
To approach real-time, precise and convenient cockpit voice control, this paper provides methods to make cockpit voice instruction to conform to the requirements of human-machine communication and voice recognition together. It emphasizes that a complete and gradate voice instruct system should stem both from the air-landing language and the prompt titles in display system. It also points out that the pronunciation of vocabularies in the instruction should be selected properly to make it appropriate to the computer voice recognizer.
出处
《航空电子技术》
2012年第3期39-43,共5页
Avionics Technology
关键词
机载语音控制
语音指令
语音识别
陆空通话
噪声
语音变异
cockpit voice control, voice instruction, voice recognition, air-landing language, noise, voice variation
