Double Q-Learning for Radiation Source Detection

Zheng Liu, Shiva Abbaszadeh

Research output: Contribution to journalArticle

Abstract

Anomalous radiation source detection in urban environments is challenging due to the complex nature of background radiation. When a suspicious area is determined, a radiation survey is usually carried out to search for anomalous radiation sources. To locate the source with high accuracy and in a short time, different survey approaches have been studied such as scanning the area with fixed survey paths and data-driven approaches that update the survey path on the fly with newly acquired measurements. In this work, we propose reinforcement learning as a data-driven approach to conduct radiation detection tasks with no human intervention. A simulated radiation environment is constructed, and a convolutional neural network-based double Q-learning algorithm is built and tested for radiation source detection tasks. Simulation results show that the double Q-learning algorithm can reliably navigate the detector and reduce the searching time by at least 44% compared with traditional uniform search methods and gradient search methods.
Original languageEnglish (US)
Number of pages19
JournalSensors
Volume19
Issue number4
DOIs
StatePublished - 2019

Fingerprint

radiation sources
learning
Learning
Radiation
radiation
background radiation
Learning algorithms
reinforcement
Background Radiation
Diptera
Reinforcement learning
gradients
scanning
detectors
Surveys and Questionnaires
simulation
Detectors
Neural networks
Scanning

Cite this

Double Q-Learning for Radiation Source Detection. / Liu, Zheng; Abbaszadeh, Shiva.

In: Sensors, Vol. 19, No. 4, 2019.

Research output: Contribution to journalArticle

Liu, Zheng ; Abbaszadeh, Shiva. / Double Q-Learning for Radiation Source Detection. In: Sensors. 2019 ; Vol. 19, No. 4.
@article{cf671cc950c242da9feb7c1e45f0b2c0,
title = "Double Q-Learning for Radiation Source Detection",
abstract = "Anomalous radiation source detection in urban environments is challenging due to the complex nature of background radiation. When a suspicious area is determined, a radiation survey is usually carried out to search for anomalous radiation sources. To locate the source with high accuracy and in a short time, different survey approaches have been studied such as scanning the area with fixed survey paths and data-driven approaches that update the survey path on the fly with newly acquired measurements. In this work, we propose reinforcement learning as a data-driven approach to conduct radiation detection tasks with no human intervention. A simulated radiation environment is constructed, and a convolutional neural network-based double Q-learning algorithm is built and tested for radiation source detection tasks. Simulation results show that the double Q-learning algorithm can reliably navigate the detector and reduce the searching time by at least 44{\%} compared with traditional uniform search methods and gradient search methods.",
author = "Zheng Liu and Shiva Abbaszadeh",
year = "2019",
doi = "10.3390/s19040960",
language = "English (US)",
volume = "19",
journal = "Sensors (Basel, Switzerland)",
issn = "1424-8220",
publisher = "Multidisciplinary Digital Publishing Institute (MDPI)",
number = "4",

}

TY - JOUR

T1 - Double Q-Learning for Radiation Source Detection

AU - Liu, Zheng

AU - Abbaszadeh, Shiva

PY - 2019

Y1 - 2019

N2 - Anomalous radiation source detection in urban environments is challenging due to the complex nature of background radiation. When a suspicious area is determined, a radiation survey is usually carried out to search for anomalous radiation sources. To locate the source with high accuracy and in a short time, different survey approaches have been studied such as scanning the area with fixed survey paths and data-driven approaches that update the survey path on the fly with newly acquired measurements. In this work, we propose reinforcement learning as a data-driven approach to conduct radiation detection tasks with no human intervention. A simulated radiation environment is constructed, and a convolutional neural network-based double Q-learning algorithm is built and tested for radiation source detection tasks. Simulation results show that the double Q-learning algorithm can reliably navigate the detector and reduce the searching time by at least 44% compared with traditional uniform search methods and gradient search methods.

AB - Anomalous radiation source detection in urban environments is challenging due to the complex nature of background radiation. When a suspicious area is determined, a radiation survey is usually carried out to search for anomalous radiation sources. To locate the source with high accuracy and in a short time, different survey approaches have been studied such as scanning the area with fixed survey paths and data-driven approaches that update the survey path on the fly with newly acquired measurements. In this work, we propose reinforcement learning as a data-driven approach to conduct radiation detection tasks with no human intervention. A simulated radiation environment is constructed, and a convolutional neural network-based double Q-learning algorithm is built and tested for radiation source detection tasks. Simulation results show that the double Q-learning algorithm can reliably navigate the detector and reduce the searching time by at least 44% compared with traditional uniform search methods and gradient search methods.

U2 - 10.3390/s19040960

DO - 10.3390/s19040960

M3 - Article

C2 - 30813497

VL - 19

JO - Sensors (Basel, Switzerland)

JF - Sensors (Basel, Switzerland)

SN - 1424-8220

IS - 4

ER -