DOE PAGES title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Double Q-Learning for Radiation Source Detection

Journal Article · · Sensors
DOI: https://doi.org/10.3390/s19040960 · OSTI ID:1496491

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.

Sponsoring Organization:
USDOE National Nuclear Security Administration (NNSA)
Grant/Contract Number:
NA0002576
OSTI ID:
1496491
Journal Information:
Sensors, Journal Name: Sensors Journal Issue: 4 Vol. 19; ISSN SENSC9; ISSN 1424-8220
Publisher:
MDPI AGCopyright Statement
Country of Publication:
Switzerland
Language:
English

References (19)

Q-learning journal May 1992
Adaptive Bayesian Sensor Motion Planning for Hazardous Source Term Reconstruction journal July 2017
Information driven search for point sources of gamma radiation journal April 2010
Human-level control through deep reinforcement learning journal February 2015
The Lost Source, Varying Backgrounds and Why Bigger May Not Be Better conference January 2002
Detection and parameter estimation of multiple radioactive sources conference July 2007
On Localisation of a Radiological Point Source conference February 2007
Networked sensing systems for detecting people carrying radioactive material
  • Chandy, Mani; Pilotto, Concetta; McLean, Ryan
  • 2008 Fifth International Conference on Networked Sensing Systems (INSS), 2008 5th International Conference on Networked Sensing Systems https://doi.org/10.1109/INSS.2008.4610916
conference June 2008
Smart radiation sensor management journal September 2008
Deep Reinforcement Learning: A Brief Survey journal November 2017
Reinforcement Learning: An Introduction journal September 1998
Efficient strategies for low-statistics nuclear searches journal June 2006
Examination of Count-Starved Gamma Spectra Using the Method of Spectral Comparison Ratios journal August 2007
Iterative Estimation of Location and Trajectory of Radioactive Sources With a Networked System of Detectors journal April 2013
Kernel-Based Machine Learning for Background Estimation of NaI Low-Count Gamma-Ray Spectra journal June 2013
Maximum Likelihood Localization of Radioactive Sources Against a Highly Fluctuating Background journal December 2015
Radiological Source Detection and Localisation Using Bayesian Techniques journal November 2009
Cooperation between an unmanned aerial vehicle and an unmanned ground vehicle in highly accurate localization of gamma radiation hotspots journal January 2018
Deep Reinforcement Learning with Double Q-Learning journal March 2016