Research And Implement Of Speech Interactive System For Space OS Services Secretary

Speech interactive system is designed so that the computer will be able to listen and speak.After 50 years' research and development,speech recognition technology has come a long way.Speech synthesis technology has also made rapid development,and has been widely used.The interactive voice technology,which based on speech recognition,speech synthesis and semantic understanding,also has been the development and application. With the development of Space OS intelligence,this paper explore the implementation of speech interactive,which is the most natural and convenient and effective way of human-computer interaction.The realization of a particular person's particular set of voice commands to operate,accurate operation,to the extent of practical use for Space OS is the goal of this research.This article first introduces the development of speech interactive, expounds the background and the content of the task,and then introduces the speech interactive system's overall design for the Space OS service Secretary,the specific use of a particular set of voice commands and adopt the decision-making speech synthesis feedback to determine the operation of voice to reach 100%accuracy,so that the study reached a practical level.And then introduces the basic principles and techniques of speech recognition,speech synthesis and Space OS which apply to this task.For aspects of speech recognition HMM and HTK are introduced;for aspects of speech synthesis the principles of formant synthesis and eSpeak are introduced.Then focus on how to apply the HMM-based speech recognition system for the entire process,including identification of the choice of primitives,feature extraction,model training and recognition,and HTK system used to generate training SpaceVoiceCMD voice command set;then elaborate on how to use Praat and espeakedit to perfect pronunciation of the speech synthesis unit;final details will be the process of integrating ATK into SpaceDesktop,which includes how use the SpaceVoiceCMD voice command sets and speech synthesis primitives to achieve speech recognition and voice synthesis function,and ultimately realize the simple speech interactive systems.By the realization of actual speech recognition system,speech synthesis system and a simple speech interactive system for the Space OS services secretary,this task eventually realizes specific voice command recognition, synthesis,confirm and implement for particular person.The work what this paper has done makes a good foundation for better human-computer interaction and the further development of Space OS intelligence.