University of Illinois Urbana-Champaign

Training iCub robot pitch detection with recurrent neural network and LSTM

Tang, Steven

This item is only available for download by members of the University of Illinois community. Students, faculty, and staff at the U of I may log in with your NetID and password to view the item. If you are trying to access an Illinois-restricted dissertation or thesis, you can request a copy through your library's Inter-Library Loan office or purchase a copy directly from ProQuest.

Permalink

https://hdl.handle.net/2142/100036

Description

Title
Training iCub robot pitch detection with recurrent neural network and LSTM
Author(s)
Tang, Steven
Contributor(s)
Levinson, Stephen E.
Issue Date
2018-05
Keyword(s)
  • Music
  • pitch detection
  • recurrent neural network
  • humanoid robot
Abstract
This thesis is designed to investigate the development of robots with the ability to learn natural language. One quality that many languages use is pitch. Many languages such as Chinese are tonal, and a difference in pitch can change the meaning of a word completely. Past research has shown that music and early language acquisition are similar and that language can be described as a special type of music. This project is the first step to teaching the robot pitch detection and music creation. This project is modeled after ways humans can detect pitch. Musicians typically gain the ability of detecting pitches through constant exposure from practicing composed songs with their instruments. The goal of this thesis is to develop pitch detection with the use of a recurrent neural network to recognize notes on an electrical keyboard and perform them back in the same order. Fast Fourier Transforms are used to preprocess the data for the recurrent neural network, and the motor library and the forward and inverse kinematics library for the iCub robot are used to move the arm of the robot to play on the keyboard. This project is split into three modules: the preprocessing module, the learning module, and the motor module. Improvements and future developments will also be discussed.
Type of Resource
text
Language
en
Permalink
http://hdl.handle.net/2142/100036

Owning Collections

Senior Theses - Electrical and Computer Engineering PRIMARY
The best of ECE undergraduate research