Font Size: a A A

Design Of Game System Based On Voice Recognition

Posted on:2014-04-20Degree:MasterType:Thesis
Country:ChinaCandidate:Y LiFull Text:PDF
GTID:2268330392471389Subject:Signal and Information Processing
Abstract/Summary:PDF Full Text Request
In recent years, voice recognition is considered as one of human-computerinteraction, and becomes a research highlight. The design of game system based onvoice recognition in my work not only catches the trend of human-computer interactionand but also finds applications in game system, my work will have a bright future.The system is based on Altera DE2-70development kit. Combining the powerfulparallel process ability of FPGA and the flexibility of Nios II soft-core processertogether, it customized a SOPC system according to different functions. My workutilizes the common voice recognition algorithm and transplants the μC/GUI graphiclibrary to create a friendly human-computer interaction window. As a result, it builds agame system based on voice recognition with stable function, friendly window, low castand easy update. My work is divided into two parts: hardware design and softwaredesign.The core job of hardware is to build a SOPC system based on Nios II soft-coreprocessor according to the Avalon bus standard. The work specifically includes:①Design voice acquisition module for digitization analog voice signal by usingWM8731and the Ping-Pong operation principle;②Write PS/2custom module and usesoftware decoding to realize keyboard communication with FPGA for facilitating theentry of customer information;③Build touch screen hardware module forhuman-computer interaction interface by combining the time sequence of LTM touchscreen and IP core. Furthermore, memory, timer and other peripheral are applied.Finally, all the hardware devices are integrated into the SOPC system by Quartus II.Software design is based on the SOPC system to realize the software function inthe system according to the target of my work. This job mainly consists of three part:①Analysis the voice recognition algorithm and select the double threshold voicecollection algorithm for valid voice extraction, MFCC parameter extraction algorithmfor feature information extraction which characterize the voice and vector quantizationalgorithm based on genic algorithm for codebook generation and recognition judgment,utilize hardware resource for clustering acceleration by taking advantage of the highparallel processing capability of FPGA;②Transplant the μC/GUI graphic library toprovide a friendly human-computer interaction interface for three system workingmodes(clients set-up mode, voice log-in mode and clients deletion mode) by applying the Chinese font, window controller, photos and dialog box;③Design games whichconsist of the embedded platform games(Sokoban and Snake, which are realized byembedded platform software combines with μC/GUI graphic library) and platformcontrolled PC games(originally installed on PC, which are realized by serial voicesignal sending from the embedded platform).From the experiment results, the system can have advantages of high recognition,friendly window, simple operation, convenient to use and so on. Furthermore, thesystem casts low, has a small size and is easy to update. These merits make the systeminto a promising product, which will be widespread used.
Keywords/Search Tags:Voice recognition, Games system, SOPC technology, μC/GUI
PDF Full Text Request
Related items