Research On SmartHome Embedded Control System Based On Speech Perception

With the development of modern science and technology,the way of human-computer interaction is also improving.From the previous manual control to mobile phone control gradually turned to the most natural,free and direct way of human voice interaction.Smart home has the advantages of comfort,health,safety and efficiency,and one of its development directions is embedded speech recognition and control.This paper designed a set of intelligent home embedded control system based on voice perception.Under the condition of confirming the family members,the system can control the home equipment through the combination of voice instructions and home environment.Improved the shortcomings of current smart home,such as no identifying controller,single decision factor and so on.After preprocessing the speech signal,the Mel frequency cepstrum coefficient eigenvector is extracted;In view of the problem of the singular array in the estimation of the parameters of the Gauss mixed model,the modified EM algorithm is used to optimize the family members;Then,Gaussian mixture model was used to fit the probability distribution of the observed symbols in the hidden state of the speech,and the speech instruction recognition simulation was performed using the hidden Markov model.Under the Linux operating system,audio collection,confirmation of family members,and recognition of voice commands are implemented.In this paper,the Tiny4412 development board of Cortex-A9 is used as the upper computer of the hardware system,and transplants the Linux operating system as the software platform,designs the ZigBee wireless sensor control network as the lower position machine.The ALSA audio library and the algorithms studied in this paper are transplanted in the upper computer to implement audio collection and storage,family member confirmation and voice command identification;The lower computer is used to collect environmental information in real time and control home facilities.After the system is started,the family member is first confirmed by voice,and then the voice instruction recognition result and the real-time environmental information jointly decide to form a control instruction.The host computer sends the control instruction to the lower computer through the serial port to control the corresponding home equipment.The system is tested and analyzed under experimental conditions.The test results show that:the system can control home facilities through the combination of voice commands andreal-time environmental information when family members are confirmed.