Research On Speech Recognition System Based On MCU

With the continuous development of computer technology, it is hoped to achieve man-computer dialogue by voice.As a result, speech recognition technology came into being. Speech recognition is a kind of technology that is using computer to transfer the voice signal to a command by identification and understanding. Speaker-depended isolated-word speech recognition is the most widely used speech recognition technology. However, the cost of hardware devices in this regard is too high at present, resulting in an unnecessary waste of resources.Given the status quo, through the studying of the characters of single-word, a Speaker-Depended Isolated-word speech recognition system is designed with the combination of MCU and other hardware device.The main research and results are summarized as follows:Firstly, 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 rate and short-term average rate is presented which enhanced the applicability for different types of words.LPCC (Linear Predictive Cepstral Coding) is extracted from the signal as the speech characteristic parameters.DTW (Dynamic Time Warping) algorithm, which is improved to reduce the amount of data in the matching process by using global constraint, is used for the matching of the model.The average training method is used to enhance the robustness of the system in the course of the training templates for the randomness of speech signal.Secondly, C8051F040 microcontroller selected as the core, circuit design of the system is shown. Then, the paper details the software components, including the concrete realization of the process of the algorithm.In addition, taking into account the characteristics of the system itself, the parameters of the algorithm are calculated.Finally, a lot of debugging and test are carried out under different noise environment and language environment. Experiment results indicate that this control system is stable, and can achieve a good recognition result for special words in quiet conditions.At the same time, due to the low cost, simple operation, the system will have a good application prospects and will be widely used in future.