Design And Development Of Embedded Speech Recognition System On Robot

As technology in embedded equioment improved,embedded equipment applicationshave been used in many aspects of people’s lives.When we interaction with them,more andmore people hope we can directly talk to these applications to tell them the tasks,just like wecomunicate with our friends.So people began to research on embedded speech recognition.Embedded speech recognition technology liberate people from the traditional keyboard,mouse, writing pad interaction ways, and help to make people’s daily life more convenient.Sofar, there are a lot algorithms to handle speech recognition, but the poor processing capacityand RAM of embedded systerm has limited their performance on enbedded system. Therefore,the research of embedded speech recognition are practical significant and has broadapplication prospects.This paper’s research work is based on ARM9platform, take use of Sphinx speechrecognition toolbox and Google voice recognition API, designed a local speech recognitionsystem that used when there is no network connection and an online speech recognitionsystem that network is avaliable.This two systems complet each other and can use in theactual situations.This paper includes the following several aspects:First，illustrate the research background,development process and reaserach status. Thenintroduces the basic knowledge of speech recognition.Second，introduce the embedded robot platform, the robot’s functions and the interface onthe robot platform.Third,illustrate the method of using PocketSphinx toolbox to build a local speechrecognition system.Forth,design an online speech recognition system with Google voice recognition API.Fifth, Compare a learning card instance with a querry weather instance,verify the speechrecognition system is correct and practical.Last，summarize the robot embedded speech recognition systerm illustrating in this paper.