Design Of Intelligent Nursing Bed Speech Recognition System

Speech recognition is a kind of technology that enables device to recognize and comprehend speech signal, it’s an important technical means for human-computer interaction. Good performance and low cost embedded Speech recognition system which providing convenient service for human-computer interaction will have broad application prospects,Through analyzes the characteristics and research status of embedded speech recognition system, a design of a low-cost small vocabulary isolated word speech-specific human recognition module was proposed. The design is to make the module has friendly man-machine interface, can work both independently and through the interface module coordination with other systems.The speech recognition implementation in embedded systems is influenced by many factors. In embedded systems environment, computing rate is relatively lower, storage space is limited, and requires certain real-time. According to the mechanism of voice production, a variety of speech recognition algorithms are introduced. After the pretreatment on the speech signal, a new algorithm based on zero crossing rates and short-term average rate is presented which enhanced the applicability for different types of words. Linear predictive campestral coding is extracted from the signal as the speech characteristic parameters. Although the speech recognition technology won some achievement in high performance platform, it’s necessary to due further analysis and research on speech recognition in embedded platform. This paper’s main research works are listed as follows:1) A low-cost Microcontroller SPCE061A was selected which is suitable for speech signal acquisition and processing. The hardware platform for speech recognition and test was designed.2) Combined the characteristics of the MCU to design the system software. The use of Surplus Technology Company to provide the function library, shorten the development cycle.3) The performance of speech recognition modules was tested. In the speech recognition module is realized on a conventional control nursing bed. Experiments show that the speech recognition module has achieved objectives of design.