Research On Voice Interaction And Control System Of Home Accompanying Robots

Recently,due to the development of deep learning,speech interactive technology has been used in many industry fields.However,in the real environment,due to the existence of interference,voice recognition is poor.Although the speech interactive software is being used widely in market,but there is a certain distance that the users are widely acceptable it in real life.China entered the aging society.Now China has accelerated the research of service robots for elderly,which is one of the most important research directions in the current robot researches.Service robots for elderly have huge market value.This paper introduces the speech control of service robot for elderly using the technology of speech separation and speech recognition.The paper makes a detailed analysis of the structure for supervised speech separation.Through the simulation experiment of single channel speech separation system based on deep learning,the different parameters are compared,and this paper discussed the objective function,the number of hidden layers and the activation function.Through comparing the experimental results,a suitable for Chinese speech separation system model is presented.The combination of Mel-Frequency Cepstral Coefficient and Relative Spectral Transform-Perceptual Linear Prediction is used as the feature of speech recognition and the simulation is carried out,which is based on Chinese isolated word in use of deep natural network.According to the experimental results,the effect of the combination on the accuracy of speech recognition is analyzed.Through designing of the speech control system for robot caregivers,this chapter uses JAVA language programming on the eclipse platform to achieve the system,and gives a detailed description of the control interface.The recognition system and hardware system are detailed described.Debugging the system of speech interactive control robot,which is proposed in this paper.The experiment verifies the performance of the interactive control system of the robot caregivers,and realizes the man-machine communication through the speech control system.In this paper,it is easy to cause inaccurate recognition for the elderly in the process of using robot caregivers,especially in the present of noise.This paper provides a speaker independent and isolated word speech recognition system model,and realizes the control of the robot.The system is also suitable for the other man-machine communication systems,such as the vehicle system.