Title: Vision-based robotics at CIMAT
Abstract: In this seminar, I will present some works that our robotics team at CIMAT (Mexico) has developed recently, and that reflect my own research interests in my sabbatical stay at IRISA. In visual perception, I will introduce our intention-aware visual pedestrian tracking and prediction algorithms for robot navigation among crowds and a technique for the estimation of the absolute scale of mono-SLAM reconstruction based on the detection of categorical visual data. In applications of control and planning, I will describe our recent works in homography-based visual control for the navigation of humanoid robots in visual memories and optimization-based planning using regression techniques.
Bio: Jean-Bernard Hayet graduated from Ecole Nationale Supérieure de Techniques Avancées in 1999. He got his master degree from University Paris VI and his Ph.D. degree from University of Toulouse in 2003. His doctoral dissertation (at LAAS) tackled the problem of landmark-based navigation with computer vision in indoor environments and was supervised by Dr. M. Devy and Dr. F. Lerasle. From 2003 to 2006 he did a post-doctoral stay in visual tracking at University of Liege with Pr. Justus Piater. Since 2007 he is researcher at the Centro de Investigacion en Matematicas, in Mexico, and is focusing his research on robot vision and visual tracking. He is member of the Mexican National System of Researchers at level 1.