Winners of the Robotdalen Scientific Award

Winner 2009 - Dr Davide Scaramuzza
Swiss Federal Institute of Technology in Switzerland

Thesis: Omnidirectional Vision: from Calibration to Robot Motion Estimation

The Jury´s statement: Scaramuzza's thesis presents three important engineering contributions, capped with a fundamental scientific advance via his algorithmic approach to the utilization of omnidirectional vision in support of environmentally-referenced navigation of mobile robots. His engineering contributions include a toolkit for the notoriously difficult problem of calibrating omnidirectional cameras, robust image feature extraction based on semantic-level descriptors of line details, and a rigorous mutual calibration scheme for camera, odometer, and rangefinder sensors. His scientific contribution -- resulting in a pending patent -- is a 1-point RANSAC algorithm that he shows to be an effective and computationally efficient reduction of the standard 5-point RANSAC algorithm when vehicle motion is non-holonomic.



Winner 2008 - Dr Sylvain Calinon,
University Ecole Polytechnique Federale de Lausanne in Switzerland

The thesis examines how a human instructor can teach manipulation skills to a robot by first demonstrating the required motions, then leading the robot through them and coaching the robot’s refinement of its skill level by providing feedback using gestures or voice emphasis.

The jury's statement: The jury was impressed by the practical and economic importance of the problem, the systematic analysis and innovative layered approach of its implementation, the completeness of the solution, and the open release of source code, demonstrating confidence in the solution’s independence from specifics of any particular robot hardware or manipulation task.

Read an interview with Sylvain Calinon

Winner 2007 - Dr Eric Demeester, Katholieke Universiteit in Belgium

Thesis: "User-adapted plan recognition and shared control for wheelchair driver assistance under uncertainty”.

The thesis deals with how robots should behave in order to obtain natural and intuitive human-robot interaction, with robotic wheelchairs as test case.

The jury’s statement: A beautifully organized and prepared application describing the candidate's research in human-robot interaction with the crucial goal of "understanding what the user had in mind", excellently implemented in the important domain of intelligent wheelchair mobility. The jury was impressed by the importance of the underlying problem and the relevance of the application domain, the extensibility of the approach, and the integration of theory, modelling, and hands-on implementation.