Optimal sensory encoding and robot perception and navigation

August 7, 2012 1:30 PM

302- 309




Consider the difficulty of an autonomous robot performing a search and rescue mission in a hazardous environment containing broken walls, rubble, fire, smoke, and fog.

How is it possible to sense and navigate in such noisy and uncertain environments? I will first review recently developed algorithms for mapping, path planning, and locomotion control, and then show how biological neurons are able to encode perceptual information in a Bayes optimal manner. In particular, I will discuss our recent results relating the Fisher Information measure to optimal encoding using generalized Cramer-Rao bounds.

This page is maintained by Yumi Yi (ymyi@bi.snu.ac.kr).
Last update: July 30, 2012