Design Of Intelligent Housekeeper Robot Based On Sphinx Speech Recognition

In recent years,with the rise of artificial intelligence,People are gradually becoming interested in how to make robots understand human languages.Language is the most directly and conveniently tool for human beings to communicate and transmit information.If this tool is also applicable to the interaction between robot and human,people's daily life will become more convenient and simple.With this concept,speech recognition technology arises at the historic moment.Based on this background,this thesis applies the research of speech recognition.By constructing the hardware and software platform of embedded system,the intelligent home housekeeper robot with speech recognition function is designed.The main contents of this paper are as follows:Firstly,this paper expounds the research background of this topic,and briefly summarizes the research status of speech recognition technology and intelligent home at home and abroad.Secondly,the basic control principle is proposed for Sphinx embedded speech recognition system.By studying the speech recognition algorithms of dynamic time regulation(DTW)and(HMM),the basic idea of HMM model algorithm is applied to the robot speech recognition system.Sphinx speech recognition system is used as the test platform to realize the transplantation of speech recognition.Then,use the Exynos4412 as the main controller to design the hardware system of the housekeeper robot.It includes: main control platform circuit,peripheral sensing control circuit,network communication circuit,video monitoring circuit and voice input circuit.At the same time,through the tailoring of embedded Linux system,the software control system of family housekeeper robot is designed.It includes the design of the embedded control system,the design of the peripheral sensing device driver,the program design of speech recognition and the design of the robot control interface program.Finally,the robot speech recognition function,home environment monitoring function and video monitoring function were tested.Among them,the system's speech recognition function,the overall recognition accuracy rate reached more than 90%,basically meet the performance requirements,while according to the test situation,this design is summarized and prospected.