Nikolay Atanasov, Feb 2015

From MurrayWiki
Revision as of 21:18, 5 February 2015 by Abaetica (Talk | contribs)

Jump to: navigation, search

Nikolay Atanasov, a student at Penn working with George Pappas and Kostas Danilidis will be visiting Caltech on 9 Feb (Mon). He will give a group meeting talk on some of his recent work. If you would like to meet with him, please sign up below:

Schedule

Monday:

  • 10:00a: Richard (109 Steele Lab)
  • 10:30a: Open
  • 11:15a: Open
  • 12p-1:15p: Group meeting presentation, 121 Annenberg
  • 1:15p: Vasu (location TBD)
  • 2:00p: Open
  • 2:45p: Open
  • 3:30p: Open
  • 4:15p: Yorie Nakahira
  • 5:00p: Done for the day

Seminar info

Active Information Acquisition with Applications in Robotics and Computer Vision
Nikolay Atanasov, University of Pennsylvania

Monday, February 9th, 12 pm
121 Annenberg

Abstract: The remarkable advances in sensing and mobility for robotic systems allow us to address some important information gathering problems such as environmental monitoring, source seeking, simultaneous localization and mapping (SLAM), and object recognition. The goal of active information acquisition is to design control strategies for multiple sensing robots, aimed at improving the accuracy and efficiency of tracking an evolving phenomenon of interest (target). First, we present a greedy approach to information acquisition via applications in source seeking and mobile robot localization. Next, we consider nonmyopic information acquisition in the case of linear Gaussian models and develop an approximate algorithm with suboptimality guarantees. The algorithm is computationally fast, since it exploits linearized sensing models, and memory efficient, since it exploits sparsity in the target model. Decentralized control and estimation are discussed and an application to active multi-robot SLAM is presented. Finally, we consider nonmyopic information acquisition with general sensing models and present applications in active object recognition.

Biography: Nikolay A. Atanasov received a B.S. degree in electrical engineering from Trinity College, Hartford, CT, in 2008 and an M.S. degree in electrical engineering from the University of Pennsylvania, Philadelphia, PA, in 2012. Currently, he is working towards the Ph.D. degree in electrical and systems engineering at the University of Pennsylvania under the supervision of Professor George J. Pappas and Professor Kostas Daniilidis. His research interests are in active information acquisition with applications to robotics and computer vision, encompassing problems such as active object recognition, active localization and mapping, environmental monitoring, security and surveillance, search and rescue, etc. The fields relevant to his research are robotics, control theory, optimization, and computer vision.