{"id":{"repo_id":"uiuc","oai_identifier":"oai:www.ideals.illinois.edu:2142/104847"},"canonical_url":"https://search.dev.ndltd.org/etd/uiuc/oai:www.ideals.illinois.edu:2142/104847","repository":{"repo_id":"uiuc","name":"University of Illinois - Urbana-Champaign","base_url":"https://www.ideals.illinois.edu/oai-pmh"},"display":{"title":"Reconstruction of urban radiation landscape using machine learning methods","abstract":"Efficiently monitoring a geographic region's radiation level and detecting anomalous radiation sources is an essential issue in homeland security. This task includes identifying illicit movement of special nuclear material, locating unusual radioactive events, and estimating the intensity of radioactive sources to name a few. Besides those anomalous radiation sources, there is naturally occurring radioactive material presented in air, soil and building materials. Radiation emitted from those materials compose the background radiation, which fluctuates in both space and time. The urban radiation landscape consists of the anomalous radiation sources and the background radiation. In this thesis, we present our work on reconstructing the urban radiation landscape using mobile sensor networks, which has two interconnected focuses. One is to model the background radiation; the other is to detect and search for anomalous radiation sources. Modeling of background radiation is conducted in two steps: retrospective modeling and prospective modeling. The retrospective modeling focuses on estimating visited positions’ radiation intensities, in which a maximum likelihood estimation method is developed to decouple and estimate temporal fluctuation and spatial distribution of background radiation. The prospective modeling focuses on predicting background radiation intensities, in which the Gaussian process regression is applied to predict unvisited positions' background spatial distributions, and recurrent neural network models are trained to predict future background temporal fluctuations. An integrated anomalous radiation source detection algorithm is developed to detect radiation sources in urban radiation landscape. Background radiation models are combined in the algorithm to eliminate false alarms produced by high background regions and temporal background fluctuations. A double Q-learning based anomalous source searching algorithm is investigated to navigate the detector searching for sources.","abstract_html":"Efficiently monitoring a geographic region's radiation level and detecting anomalous radiation sources is an essential issue in homeland security. This task includes identifying illicit movement of special nuclear material, locating unusual radioactive events, and estimating the intensity of radioactive sources to name a few. Besides those anomalous radiation sources, there is naturally occurring radioactive material presented in air, soil and building materials. Radiation emitted from those materials compose the background radiation, which fluctuates in both space and time. The urban radiation landscape consists of the anomalous radiation sources and the background radiation. In this thesis, we present our work on reconstructing the urban radiation landscape using mobile sensor networks, which has two interconnected focuses. One is to model the background radiation; the other is to detect and search for anomalous radiation sources. Modeling of background radiation is conducted in two steps: retrospective modeling and prospective modeling. The retrospective modeling focuses on estimating visited positions’ radiation intensities, in which a maximum likelihood estimation method is developed to decouple and estimate temporal fluctuation and spatial distribution of background radiation. The prospective modeling focuses on predicting background radiation intensities, in which the Gaussian process regression is applied to predict unvisited positions' background spatial distributions, and recurrent neural network models are trained to predict future background temporal fluctuations. An integrated anomalous radiation source detection algorithm is developed to detect radiation sources in urban radiation landscape. Background radiation models are combined in the algorithm to eliminate false alarms produced by high background regions and temporal background fluctuations. A double Q-learning based anomalous source searching algorithm is investigated to navigate the detector searching for sources.","abstract_has_math":false,"creators":["Liu, Zheng"],"institution":"University of Illinois at Urbana-Champaign","degree_name":"Ph.D.","degree_level":"Dissertation","degree_discipline":"Nuclear, Plasma, Radiolgc Engr","degree_department":null,"school":null,"contributors":["Kozlowski, Tomasz","Abbaszadeh, Shiva","Uddin, Rizwan","He, Niao","Sullivan, Clair Julia"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2019,"date_issued":"2019-08-23T19:55:40Z","date_published":"2019-08-23T19:55:40Z","updated_at":"2026-07-22T22:24:42Z","subjects":["radiation detection","mobile detector","sensor network","anomalous radiation source detection","reinforcement learning","machine learning","maximum likelihood estimation"],"languages":["eng"],"rights":["Copyright 2019 Zheng Liu"],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"http://hdl.handle.net/2142/104847","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Kozlowski, Tomasz","Abbaszadeh, Shiva","Uddin, Rizwan","He, Niao","Sullivan, Clair Julia"]},{"key":"dc:creator","label":"Author","values":["Liu, Zheng"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2019-08-23T19:55:40Z","2019-04-17","2019-05"]},{"key":"dc:type","label":"Dc Type","values":["text"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Nuclear, Plasma, Radiolgc Engr"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Dissertation"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Ph.D."]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["University of Illinois at Urbana-Champaign"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["radiation detection","mobile detector","sensor network","anomalous radiation source detection","reinforcement learning","machine learning","maximum likelihood estimation"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["eng"]},{"key":"dc:rights","label":"Dc Rights","values":["Copyright 2019 Zheng Liu"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["http://hdl.handle.net/2142/104847"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Efficiently monitoring a geographic region's radiation level and detecting anomalous radiation sources is an essential issue in homeland security. This task includes identifying illicit movement of special nuclear material, locating unusual radioactive events, and estimating the intensity of radioactive sources to name a few. Besides those anomalous radiation sources, there is naturally occurring radioactive material presented in air, soil and building materials. Radiation emitted from those materials compose the background radiation, which fluctuates in both space and time. The urban radiation landscape consists of the anomalous radiation sources and the background radiation. In this thesis, we present our work on reconstructing the urban radiation landscape using mobile sensor networks, which has two interconnected focuses. One is to model the background radiation; the other is to detect and search for anomalous radiation sources. Modeling of background radiation is conducted in two steps: retrospective modeling and prospective modeling. The retrospective modeling focuses on estimating visited positions’ radiation intensities, in which a maximum likelihood estimation method is developed to decouple and estimate temporal fluctuation and spatial distribution of background radiation. The prospective modeling focuses on predicting background radiation intensities, in which the Gaussian process regression is applied to predict unvisited positions' background spatial distributions, and recurrent neural network models are trained to predict future background temporal fluctuations. An integrated anomalous radiation source detection algorithm is developed to detect radiation sources in urban radiation landscape. Background radiation models are combined in the algorithm to eliminate false alarms produced by high background regions and temporal background fluctuations. A double Q-learning based anomalous source searching algorithm is investigated to navigate the detector searching for sources.","Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2019-08-22 without embargo terms","The student, Zheng Liu, accepted the attached license on 2019-04-16 at 14:51.","The student, Zheng Liu, submitted this Dissertation for approval on 2019-04-16 at 15:03.","This Dissertation was approved for publication on 2019-04-17 at 09:59.","DSpace SAF Submission Ingestion Package generated from Vireo submission #13681 on 2019-08-22 at 14:44:09","Made available in DSpace on 2019-08-23T19:55:40Z (GMT). No. of bitstreams: 2 LIU-DISSERTATION-2019.pdf: 27033832 bytes, checksum: b83953eff82e6b025fa3f3bca92da760 (MD5) LICENSE.txt: 4206 bytes, checksum: 3e7cf70cac5544f920b3f24fad6c81b0 (MD5) Previous issue date: 2019-04-17"]},{"key":"dc:format","label":"Dc Format","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["Reconstruction of urban radiation landscape using machine learning methods"]}]}],"canonical_facts":{"dc:contributor":["Kozlowski, Tomasz","Abbaszadeh, Shiva","Uddin, Rizwan","He, Niao","Sullivan, Clair Julia"],"dc:creator":["Liu, Zheng"],"dc:date":["2019-08-23T19:55:40Z","2019-04-17","2019-05"],"dc:description":["Efficiently monitoring a geographic region's radiation level and detecting anomalous radiation sources is an essential issue in homeland security. This task includes identifying illicit movement of special nuclear material, locating unusual radioactive events, and estimating the intensity of radioactive sources to name a few. Besides those anomalous radiation sources, there is naturally occurring radioactive material presented in air, soil and building materials. Radiation emitted from those materials compose the background radiation, which fluctuates in both space and time. The urban radiation landscape consists of the anomalous radiation sources and the background radiation. In this thesis, we present our work on reconstructing the urban radiation landscape using mobile sensor networks, which has two interconnected focuses. One is to model the background radiation; the other is to detect and search for anomalous radiation sources. Modeling of background radiation is conducted in two steps: retrospective modeling and prospective modeling. The retrospective modeling focuses on estimating visited positions’ radiation intensities, in which a maximum likelihood estimation method is developed to decouple and estimate temporal fluctuation and spatial distribution of background radiation. The prospective modeling focuses on predicting background radiation intensities, in which the Gaussian process regression is applied to predict unvisited positions' background spatial distributions, and recurrent neural network models are trained to predict future background temporal fluctuations. An integrated anomalous radiation source detection algorithm is developed to detect radiation sources in urban radiation landscape. Background radiation models are combined in the algorithm to eliminate false alarms produced by high background regions and temporal background fluctuations. A double Q-learning based anomalous source searching algorithm is investigated to navigate the detector searching for sources.","Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2019-08-22 without embargo terms","The student, Zheng Liu, accepted the attached license on 2019-04-16 at 14:51.","The student, Zheng Liu, submitted this Dissertation for approval on 2019-04-16 at 15:03.","This Dissertation was approved for publication on 2019-04-17 at 09:59.","DSpace SAF Submission Ingestion Package generated from Vireo submission #13681 on 2019-08-22 at 14:44:09","Made available in DSpace on 2019-08-23T19:55:40Z (GMT). No. of bitstreams: 2 LIU-DISSERTATION-2019.pdf: 27033832 bytes, checksum: b83953eff82e6b025fa3f3bca92da760 (MD5) LICENSE.txt: 4206 bytes, checksum: 3e7cf70cac5544f920b3f24fad6c81b0 (MD5) Previous issue date: 2019-04-17"],"dc:format":["application/pdf"],"dc:identifier":["http://hdl.handle.net/2142/104847"],"dc:language":["eng"],"dc:rights":["Copyright 2019 Zheng Liu"],"dc:subject":["radiation detection","mobile detector","sensor network","anomalous radiation source detection","reinforcement learning","machine learning","maximum likelihood estimation"],"dc:title":["Reconstruction of urban radiation landscape using machine learning methods"],"dc:type":["text"],"thesis:degree_discipline":["Nuclear, Plasma, Radiolgc Engr"],"thesis:degree_level":["Dissertation"],"thesis:degree_name":["Ph.D."],"thesis:institution_name":["University of Illinois at Urbana-Champaign"]},"updated_at":"2026-07-22T22:24:42Z"}