Research and Application of Computer Artificial Intelligence Technology in Machine Pronunciation

Abstract

The traditional oral English training system uses a layered training strategy, which cannot meet the system performance design requirements. For this reason, the thesis designs a computer artificial intelligence system for oral English training based on human-computer interaction. The system hardware structure is composed of C8051F021 single chip microcomputer, LM317 chip, LM337 chip, A/D chip and IN5822 diode, which is used to reliably control the sampling frequency of the chip. At the same time, the paper introduces Sugeno points to realize the correct evaluation of spoken pronunciation. First, the randomness of natural language pronunciation and the instability of the speech processing system are modeled by fuzzy measure and credibility, and then they are integrated into the Sugeno integration framework to evaluate the language score instead of the specific score. The experimental results prove the credibility and stability of the evaluation model in closed and open tests.