Christian Barat

About

Christian Barat is a leading researcher in autonomous underwater robotics and low-cost perception systems, with a career focused on enabling robots to navigate and understand complex environments through innovative sensor fusion. His most significant contributions center on autonomous benthic boundary tracking, where he developed pioneering algorithms that allow an Autonomous Underwater Vehicle (AUV) equipped with a mechanically scanning profiler sonar to detect and follow transitions between different seafloor regions in real time. This work, detailed in his most-cited paper (25 citations), established a classical control loop closed around sonar data, a foundational technique for environmental monitoring and seafloor mapping. Barat has also made substantial advances in multi-modal sensor integration, notably in automatic sonar-to-video calibration (13 citations) by exploiting natural contours, enabling precise coregistration of underwater imagery and range data. His research extends to assistive robotics, where he has developed robust stereoscopic vision systems for poor-quality cameras (14 citations) and low-cost ultrasonic localization for mobile robots aiding disabled individuals. Barat’s work demonstrates that effective autonomy can be achieved even with inexpensive, low-metrological-quality sensors through clever signal processing and statistical classification, bridging neural and statistical approaches. His contributions remain influential in underwater robotics and service robotics communities.

Research Focus

Key Achievements

Top Papers

1. 1
   Benthic boundary tracking using a profiler sonar

   25 citations · 2004
2. 2
   Matching segments in stereoscopic vision

   14 citations · 2001
3. 3
   Exploiting natural contours for automatic sonar-to-video calibration

   13 citations · 2005
4. 4
   Off-line localisation of a mobile robot using ultrasonic measurements

   9 citations · 2000
5. 5
   Tracking benthic boundaries using a profiler sonar: a mixture model approach

   4 citations · 2003
6. 6
   Neural and statistical classifiers. Can such approaches be complementary?

   4 citations · 2002
7. 7
   Toward a versatile ultrasonic sensor

   3 citations · 2002
8. 8
   Active high dynamic range mapping for dense visual SLAM

   3 citations · 2017
9. 9
   Distinction between objects with ultrasonic biaural system and only amplitude

   3 citations · 2002

Key Collaborators

Contact & Links