中文语音识别技术在塔台模拟机中的研究与应用  

The Design and Implementation of Chinese speech recognition function in tower simulator

作　　者：杨璐宇[1] 

机构地区：[1]昆明船舶设备研究试验中心,云南昆明650011

出　　处：《微计算机信息》2012年第10期243-245,共3页Control & Automation

摘　　要：塔台模拟机是一个大型仿真空管模拟训练系统,其使用的陆空通话标准具有语法结构单一,通话语音标准,通话范围固定等特点。结合这些特点对如何将现代语音识别技术集成到系统中进行了研究与试验,并给出了以SAPI为开发基础的相关实现。结果显示,基于SAPI的中文语音识别技术在系统中工作识别率高,识别速度快,确实能达到减轻空管员在训练中因人员短缺产生的训练负担、提高训练效率目的。Tower simulator is a kind of large scale simulation ATC training system. It uses RTF which has the standard grammatical structure, standard pronunciation, fixed range of call. With these features modern speech recognition technologies has been integrated into the Tower simulator to research how to construct the virtual captain seats. The results showed that Chinese speech recognition technology based SAPI had better recognition rate and faster recognition speed. This method can really achieve the purpose to reduce air traffic controllers＂ training burden during training , and improve training efficiency.

关 键 词：语音识别 塔台模拟机 SAPI 上下文无关文法 模拟仿真 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]