航管模拟器的语音识别与合成技术实现

摘要

With the rapid development of the air transportation industry,it is urgent to train the air control personnel by the equipment of radar simulator and tower simulator so as to improve the handling capacity of air traffic control personnel.In order to improve the efficiency of simulation training,the speech recognition and synthesis technology in the simulator has been successfully realized based on the Microsoft speech engine Windows Speech SDK.In speech recognition,voice sets and grammar sets are added into the xml to improve the recognition rate based on basic speech recognition;a training system that dynamically generates configuration file has been designed and built through a specific word voice training to generate personal xml file.When officially using the system,the xml file must be imported into the system for reducing manual operation.In speech synthesis,the basic reading function has been realized through Microsoft engine,and the new function has been implemented that makes the voice more reality.In view of the seriousness of air traffic control training,the function of reading by specific person is realized by recording and playback.The simulation results show that the designed ATC training system has met the requirements of the current air traffic control training and improved the efficiency of the simulator.%随着航空运输业的快速发展,亟需雷达模拟机、塔台模拟机等设备培训空中管制人员,以提高空管人员的处置能力.为了提高航管雷达模拟训练的效率,基于微软语音引擎Windows Speech SDK,成功实现了航管模拟器中的语音识别与合成技术.语音识别中,基于基本语音识别,通过在xml中添加语音集和语法集来提高识别率;设计并构建了动态产生配置文件的训练系统,该系统通过特定词语的语音训练生成个人xml文件,正式使用时只需将xml文件导入以减少人工操作.语音合成中,通过微软引擎实现了基本的朗读功能,并提出了使声音更接近人声的方法.鉴于空管训练的严肃性,采用录音和播放方式实现了由特定人朗读的语音合成功能.运行情况表明,所设计构建的空管训练系统较好地满足了当前空中管制训练的要求,可明显提高航管模拟器的效率.