Design And Implementation Of Speaker-Independent Isolated-Word Speech Recognition System Based On FPGA

The technology of speech recognition has been widely used on non-contacting system, where one of the important research directions is the isolated-word speech recognition system based on the digital platform. In this thesis, the Speaker-Independent Isolated-Word Speech-Recognition system, which is based on the FPGA platform, has been researched detailedly in security system.On the analysis of feature extraction and recognition algorithms of the isolated-word, this thesis mainly focuses on the implementation of MFCC feature parameters extraction and DTW algorithms based on the FPGA. By adopting FPGA development platform based on DE2and the Verilog HDL for algorithms and controls logic, the logic modules consisting of pre-emphasis, sub-frame window, endpoint detection, MFCC feature parameters extraction and DTW pattern recognition for the voice signal have been implemented. Moreover, by combining with NIOS II processor, the isolated-word speech recognition system has been realized.Ten specific-isolated words of independent speakers are tested and analyzed. From the results, it can be seen that the system has strong instantaneity and high recognition rate.