iCub Audio Attention

An audio processing pipeline for the iCub humanoid robot.

Description

iCub Audio Attention is a research project I worked on during my Undergraduate degree. I helped to develop this project with the Cognitive Robotics Lab at the University of Lethbridge, led by Prof. Matthew Tata.

The project aims to develop signal processing algorithms and audio capture tools for auditory perception for the iCub humanoid robot.

I have received funding for this project from NSERC (USRA), Mitacs (Globalink), and the University of Lethbridge (Chinook). During the Fall 2018 semester, I was a visiting researcher at the Italian Institute of Technology in Genoa, Italy, to further develop the pipeline and ensure it works well on the iCub robot.

The project is open-source and can be found here.

We have two publications describing the main architecture, and a speech-specific filtering technique:

• A Bayesian system for noise-robust binaural sound localisation for humanoid robots
• Speech envelope dynamics for noise-robust auditory scene analysis in robotics