Question answering system of a specific domain and its implementation on a NAO humanoid robot
Liu, Siheng (2018)
Liu, Siheng
Vaasan ammattikorkeakoulu
2018
All rights reserved
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-201805097235
https://urn.fi/URN:NBN:fi:amk-201805097235
Tiivistelmä
This thesis introduces the development of a question-answer system in a specific domain and its implementation on the Nao robot. The goal of this paper is to build a limited domain, high-performance and speech supported automatic question-answering system based on NAO, the humanoid robot for hospital information.
This thesis is divided into four parts, web information extraction, the construction of the knowledge base, natural language processing and speech recognition. The data used for building a knowledge base is gathered from the official website of the hospitals by using web crawler. Then the data is stored in a database and converted to RDF type. In the second part, the questions which are asked by the users need to be converted into SPARQL to search the answer from the knowledge base. The last part is connected with the Nao robot, a recorder on Nao is used to record the voice of the users and the Google Cloud Speech API is used in this thesis for speech recognition to transfer the voice file into text. After getting the answer, the speech is synthesized by the Nao robot to interact with the user.
This thesis has achieved to use the Nao robot to answer users’ question about the some aspects of hospitals in Finland. The location, telephone number, open time can be queried.
Python is the programming language used to develop this project. The code testing is done under macOS 10.13.1, Python version 2.7, and Nao robot vision 5.
This thesis is divided into four parts, web information extraction, the construction of the knowledge base, natural language processing and speech recognition. The data used for building a knowledge base is gathered from the official website of the hospitals by using web crawler. Then the data is stored in a database and converted to RDF type. In the second part, the questions which are asked by the users need to be converted into SPARQL to search the answer from the knowledge base. The last part is connected with the Nao robot, a recorder on Nao is used to record the voice of the users and the Google Cloud Speech API is used in this thesis for speech recognition to transfer the voice file into text. After getting the answer, the speech is synthesized by the Nao robot to interact with the user.
This thesis has achieved to use the Nao robot to answer users’ question about the some aspects of hospitals in Finland. The location, telephone number, open time can be queried.
Python is the programming language used to develop this project. The code testing is done under macOS 10.13.1, Python version 2.7, and Nao robot vision 5.