[Research] Auton Lab talk: 10am on Tuesday 2/26, Gates Hall 6501

[Artur Dubrawski]

Dear Autonians,

Please come and see Prateek talk about his recent work!

Cheers
Artur

-----

Hello,

I will be giving my speaking qualifier talk Tuesday 2/26 at 10AM in GHC
6501. Everyone is invited to attend. Come for plenty of free food and
the chance to learn about the latest in Robotic and Machine Learning
Algorithms for Mobile Radiation Sensors.

Thanks,
Prateek Tandon

Title
------
Detection of Radioactive Sources using Bayesian Aggregation of Data from
Mobile Spectrometers

Time and Location
------
10AM, 26th February
GHC 6501

Abstract
------
Mobile radiation detector systems aim to help law enforcement officers
effectively identify the key sources of radiation in environments (such
as urban scenes) while minimizing false alarms caused by non-threatening
nuisance sources. In this study, we develop a Bayesian Aggregation (BA)
method for spatially aggregating evidence from multiple observations to
locate and infer properties of threatening radiation sources.

Our framework uses simulation and non-parametric techniques to model the
null and alternative distributions for detection scores of real
background measurements, either with or without simulated source
injection of threatening point sources. Information from multiple
measurements is combined with Bayesian aggregation, which aggregates
evidence by modeling multiple measurements as conditionally independent
given the underlying hypothesis about the source. Our realistic
framework supports modeling of parameters such as source intensity (to
enable detection of faint sources) and source type (to enable detection
of different sources). We compare different methods of using this
information in BA modeling and benchmark our methods against existing
methods for aggregating observations to detect sources.

Qualifier Committee
------
Prof. Artur Dubrawski
Prof. Manuela Veloso
Prof. Paul Scerri
Debadeepta Dey