Speech semantic recognition system for an assistive robotic application

摘要

This project creates a speech semantic recognition system that can be applied in the assistive robotic application by using the meaning of speech for people who suffer a permanent disability that cannot move around normally. The user interface of this speech semantic recognition system of this project are capable to receive the speech input of the user and an application interface transfers the input content of the user to an application. This speech semantic recognition system consists of input part that is speech signal. Based on the input of speech signal, the speech recognizer is used to recognize a corresponding word, the word semantic model representing the connection between a semantic database with the meaning of a word and a registered word belonging to a speech recognizer. The speech signal is recognized the word by using speech recognizer then is converted to the corresponding word-semantic model by the feature extraction approach. The converted word-semantic model is stored in the database and trained. The speech recognizer will recognize the speech signal again and notified to the application via comparing with the word-semantic model stored in the database based on feature matching method before send to the application interface.