EARS demo at Summer School

A presentation of the EARS project entitled “Signal Processing for Robot Audition” has been given by Imperial College on July 4, 2016 as part of a one-week Summer School. The two-hour presentation included also a demo and addressed students (aged 17) particularly from schools not normally sending students to engineering universities.

Demo booth at RoboCup 2016

The EARS project was presented by a booth at the RoboCup in Leipzig. The following demos have been shown from July 1-4:

  • speech dialog system and sound recognition (Aldebaran Robotics)
  • learning internal body representations for ego-noise prediction (Humboldt University Berlin)

Tutorials at Aldebaran Robotics

A tutorial on Embodied Audition for Robots was presented by Heinrich Löllmann (FAU Erlangen-Nuremberg), Christine Evers (Imperial College London) and Radu Horaud (INRIA, Grenoble) at Aldebaran Robotics in Paris on November 26, 2015:

A tutorial on variational Bayesian models and methods was given by Sileye Ba (INRIA, Grenoble) on November 27, 2015:

A presentation entitled A Variational EM Algorithm for the Separation of Moving Sound Sources was given by Laurent Girin (INRIA, Grenoble) on November 27, 2015:

RoCKIn Challenge 2015

The EARS@HOME team won the Functionality Benchmark FBM3 “Speech Understanding” of the ROCKIN@HOME Challenge, which took place in Lisbon from Nov.19-23, 2015.

Team members:

  • Heiner Löllmann (Team Leader, FAU Erlangen-Nürnberg)
  • Jean Testard (Aldebaran Robotics)
  • Vladi Tourbabin (Ben-Gurion University of the Negev)
  • Sasa Bodiroza (Humboldt Universität zu Berlin)
  • Class-Normen Ritter (Humboldt Universität zu Berlin)

Team Description Paper: TDP_EARS@HOME_ROCKIN_Challenge_2015

EARS@HOME_Team

rockin2015_60_1

 

Demo booth at IROS

The EARS project was presented by a booth at the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) in Hamburg from Sept. 29 until Oct. 2, 2015. Different demos have been shown:

  • speech dialog system and sound recognition (Aldebaran Robotics)
  • distributed architecture for interacting with Nao (INRIA)
  • learning internal body representations for ego-noise prediction (Humboldt University Berlin)
  • dictionary-based ego-noise suppression (FAU Erlangen-Nuremberg).

Demo at EuroPython

Aldebaran demonstrated a new sound event recognition system developed within the EARS project at a poster session of the EuroPhython in Bilbao from July 20-26, 2015. A video demonstrating the performance of this system can be found here.

ALD_EuroPython_2015

Royal Society Summer Science Exhibition

The exhibition Sound Interactions was presented by the Imperial College London at the Royal Society Summer Exhibition, which took place from June 30 – July 5, 2015 in London. The Summer Science Exhibition is an annual display of cutting-edge science and technology in the UK and featured 22 exhibits in 2015. A part of the exhibition Sound Interactions was a sound localization system implemented on the NAO robot. In addition, a speech recognition for variable room acoustics was shown.

2015-06-25 15.23.48

IMG_2032

EUSIPCO Tutorial on Embodied Audition for Robots

A Tutorial on Embodied Audition for Robots was presented by Heinrich Löllmann (FAU Erlangen-Nuremberg), Christine Evers (Imperial College London) and Radu Horaud (INRIA, Grenoble) at the EUSIPCO 2015 in Nice on August 31, 2015.

Presentations:
EUSIPCO_Tutorial_Robot_Audition_Introduction_HL
EUSIPCO_Tutorial_Signal_Enhancement_HL
EUSIPCO_Tutorial_Microphone_Array_Design_and_Beamforming_HL
EUSIPCO_Tutorial_Audio-Visual-Analysis_RH
EUSIPCO_Tutorial_Localisation_CE
EUSIPCO_Tutorial_Tracking_CE

Demo:
In addition to the Tutorial, a demo on ego-noise suppression for the Nao robot was presented to the participants.
EARS_Poster_demo_EUSIPCO15_20150903_AD

EARS_Demo_at_EUSIPCO_2015_20150903