The high-speed rail traction power supply dispatching simulation platform that can simulate real work scenarios is of great significance to the training of power supply dispatchers.One of the main daily tasks of traction power supply dispatchers is to make and receive calls,understand and verify the operation status of traction power supply system through the phone,and issue corresponding commands,while the current simulation platform simulation of this process is limited to the form of text.In this thesis,the calling voice of the dispatching telephone on the platform is simulated by speech synthesis,and the key content of voice reply is identified and judged,which is of great significance for improving the sense of participation and training effect of dispatch trainers.This thesis introduces the working mechanism of the dispatching telephone in the power dispatching simulation platform,and points out its shortcomings and the simulation effects that need to be achieved.According to the characteristics of the platform,the design requirements of speech synthesis system and speech recognition system are put forward.The processing of linguistics includes text regularity and word segmentation.The processing of phonetics mainly studies the rules of cooperative pronunciation.The mechanism and characteristics of three speech synthesis technologies based on physical sounding mechanism,parameter synthesis,and waveform splicing synthesis are introduced,and the waveform splicing technology is used to realize speech synthesis.A speech synthesis system for traction power supply dispatching simulation platform is developed based on VC + + 6.0,and the function of speech speed change is added.Before the development of the speech synthesis system,two auxiliary tools are developed:one is a speech editing tool for the secondary processing of the sound source files,which aims to establish the synthetic speech database.The other is thesaurus editing tool,which can group words according to the special vocabulary of high-speed railway traction power supply dispatching,establish the connection between text and syllable,and adjust the prosody of voice.Through the speech synthesis test of some texts with traction power supply dispatching specificity,the synthesized speech is natural and fluent,which verifies the effectiveness of the system and can be used for the voice simulation of incoming calls in the simulation platform.Based on the theory of speech recognition technology and VC + + 6.0,a speech recognition system of traction power supply dispatching special words is developed.In this thesis,the visual analysis and calculation of speech waveform are carried out,and the energy change trend of voiced and unvoiced segments are proposed as recognition features,which are expressed in the form of string.Feature string matching is used in speech pattern recognition.Some special vocabulary for traction power supply dispatching are selected and tested,the results show that the system can recognize the words to be tested.This system can basically realize the recognition of words in the special vocabulary for traction power dispatching. |