Speech recognition is one of the most indispensable technologies to realize intelligentized human-computer interaction, so it is a very promising direction of computer technology. After decades of research on speech recognition, great achievements have been made and part of them have already been applied in people’s daily life. The technologies of small or medium vocabulary speech recognition, such as command recognition, have become the main direction of embedded speech during the past few years and made great progress in practical system. Linux operating system have been widely used both in PC field and embedded field. With multi-platform support, Linux has been the most popular operating system in embedded field. Speech recognition should take advantage of embedded Linux to accelerate its development, so research on high performance low price speech recognition applicaiton in Linux is also very important. The main work of this paper is list as follows: 1.Summarized the main techniques of speech recognition system optimization; 2.Transplanted the human-computer speech interaction platform to Linux. First the GUI design technology and tools of Linux were introduced, and then analysed the structure, characteristics and implementation mechanism of a Windows version of human-computer interaction platform. After that a plan to implement the Linux version of human-computer interaction platform was proposed. Then the main work of transplanting this platform to Linux was completed and achieved a good result after the evaluation. 3.Implemented a dialog system in embedded system. The technology of developing and debugging program in embedded system was introduced. In this dialog system, rule match was adopted to manage content and “garbage word” model was the rejection method.
修改评论