Jean Rouat

About

Jean Rouat is a pioneering researcher in robot audition, acoustic signal processing, and multimodal artificial intelligence, whose work has fundamentally advanced how machines perceive and interpret sound in real-world environments. Based at the Université de Sherbrooke, Rouat has dedicated much of his career to solving the formidable challenge of enabling mobile robots to localize, separate, and recognize simultaneous sound sources — a problem far more complex than its human counterpart might suggest. His most influential contributions include developing robust microphone array techniques for sound source localization on mobile robots, accumulating nearly 400 citations, and pioneering beamforming combined with particle filtering for tracking moving sources in real time (272 citations). These methods gave robots a functional auditory sense capable of complementing vision without requiring line-of-sight. Rouat further advanced the field by integrating geometric source separation with post-filtering and missing feature theory to achieve reliable speech recognition even amid competing speakers — a landmark achievement demonstrated through real-time recognition of three simultaneous sentences. Beyond acoustics, Rouat co-developed HoME, a large-scale household multimodal simulation environment for training AI agents across vision, audio, and physical interaction, reflecting his broader commitment to embodied intelligence. With over 1,200 cumulative citations, his body of work remains essential reading for researchers in robotics, computational audition, and machine learning.

Research Focus

Key Achievements

Top Papers

1. 1
   Robust sound source localization using a microphone array on a mobile robot

   387 citations · 2004
2. 2
   Robust localization and tracking of simultaneous moving sound sources using beamforming and particle filtering

   272 citations · 2007
3. 3
   Enhanced robot audition based on microphone array source separation with post-filter

   136 citations · 2005
4. 4
   Localization of simultaneous moving sound sources for mobile robot using a frequency- domain steered beamformer approach

   117 citations · 2004
5. 5
   Robust Recognition of Simultaneous Speech by a Mobile Robot

   81 citations · 2007
6. 6
   HoME: a Household Multimodal Environment

   80 citations · 2017
7. 7
   Enhanced Robot Speech Recognition Based on Microphone Array Source Separation and Missing Feature Theory

   69 citations · 2006
8. 8
   Making a robot recognize three simultaneous sentences in real-time

   33 citations · 2005
9. 9
   Microphone array post-filter for separation of simultaneous non-stationary sources

   31 citations · 2004

Key Collaborators

Contact & Links