Radiation source detection from mobile sensor networks using principal component analysis

Detecting the presence of possible illicit radioactive materials in large areas is challenging because of changing background radiation, shielding effects and short collection time, especially when the radioactive materials are moving. The concept of mobile sensor networks is put forward to solve this problem. In this thesis, a small mobile sensor network is established using commercially available radiation detectors and cell phones. A spectrum decomposition and reconstruction method based on Principal Component Analysis (PCA) is proposed to work with mobile sensor networks. Two experiments are designed to test this method's performance on real-world data. The PCA-based method's performance is analyzed using receiver operating characteristic, or ROC curves. Further study finds that although the PCA-based method doesn't work well on current mobile sensor networks, its performance can be improved by increasing the radiation spectral quality.

Name thesis:degree_name

M.S.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Nuclear, Plasma, Radiolgc Engr

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2016

Author dc:creator

• Zhao, Jifu

Contributors dc:contributor

• Sullivan, Clair J.
• Mohaghegh, Zahra

• Radiation detection
• mobile sensor networks
• principal component analysis

Statement dc:rights

• Copyright 2016 Jifu Zhao

Language dc:language

en

Handle dc:identifier

http://hdl.handle.net/2142/92861

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/92861