Special Colloquium-Visibility based sensor placement for learning a priori unknown environments
Mon, 27 August, 2018
5:10pm
Speaker: Yen-Hsi Richard Tsai, Department of Mathematics, Center for Numerical Analysis, Institute for Computational Engineering and Sciences, University of Texas at Austin
Date and Time: August 27, 2018, 1:10pm-2:10pm
Place: Rome 204
Title: Visibility based sensor placement for learning a priori unknown environments
Abstract: In this talk, I will discuss the problem of sensor placement.
The question that we try to address is: where should one observe a given environment in order to learn/reconstruct the complex terrain contained in it?
The problem is motivated by the increasing availability and applications UAVs (unmanned aerial vehicles) that are equipped with sensors such as LiDARs, which provide shortest distance to obstacles from the sensors.
The essential question that we try to address is that of optimality: given an environment containing complicated three dimensional geometrical objects, what is the minimal set of “viewing” positions that allows a single sensor to survey the given environment?
I will discuss prior work and our recent work involving the use of a novel machine learning approach.
Abstract: In this talk, I will discuss the problem of sensor placement.
The question that we try to address is: where should one observe a given environment in order to learn/reconstruct the complex terrain contained in it?
The problem is motivated by the increasing availability and applications UAVs (unmanned aerial vehicles) that are equipped with sensors such as LiDARs, which provide shortest distance to obstacles from the sensors.
The essential question that we try to address is that of optimality: given an environment containing complicated three dimensional geometrical objects, what is the minimal set of “viewing” positions that allows a single sensor to survey the given environment?
I will discuss prior work and our recent work involving the use of a novel machine learning approach.