{"id":{"repo_id":"uiuc","oai_identifier":"oai:www.ideals.illinois.edu:2142/129955"},"canonical_url":"https://search.dev.ndltd.org/etd/uiuc/oai:www.ideals.illinois.edu:2142/129955","repository":{"repo_id":"uiuc","name":"University of Illinois - Urbana-Champaign","base_url":"https://www.ideals.illinois.edu/oai-pmh"},"display":{"title":"Reinforcement learning and optimization methods in sensor networks","abstract":"Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2025-10-20 without embargo terms","abstract_html":"Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2025-10-20 without embargo terms","abstract_has_math":false,"creators":["Muthuveeru-Subramaniam, Adarsh"],"institution":"University of Illinois Urbana-Champaign","degree_name":"Ph.D.","degree_level":"Dissertation","degree_discipline":"Electrical & Computer Engr","degree_department":null,"school":null,"contributors":["Veeravalli, Venugopal V.","Veeravalli, Venugopal V","Moulin, Pierre","Varshney, Lav R","Chatterjee, Sabyasachi","Zachary Hare, James"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2025,"date_issued":"2025-07-18","date_published":"2025-07-18","updated_at":"2026-07-22T22:25:06Z","subjects":["Reinforcement Learning","Optimization","Sensor Networks"],"languages":["en","eng"],"rights":["Copyright 2025 Adarsh Muthuveeru-Subramaniam"],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://hdl.handle.net/2142/129955","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Veeravalli, Venugopal V.","Veeravalli, Venugopal V","Moulin, Pierre","Varshney, Lav R","Chatterjee, Sabyasachi","Zachary Hare, James"]},{"key":"dc:creator","label":"Author","values":["Muthuveeru-Subramaniam, Adarsh"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2025-07-18","2025-08"]},{"key":"dc:type","label":"Dc Type","values":["text"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Electrical & Computer 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 Urbana-Champaign"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Reinforcement Learning","Optimization","Sensor Networks"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["en","eng"]},{"key":"dc:rights","label":"Dc Rights","values":["Copyright 2025 Adarsh Muthuveeru-Subramaniam"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["https://hdl.handle.net/2142/129955"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2025-10-20 without embargo terms","The student, Adarsh Muthuveeru-Subramaniam, accepted the attached license on 2025-07-16 at 19:42.","The student, Adarsh Muthuveeru-Subramaniam, submitted this Dissertation for approval on 2025-07-16 at 19:48.","This Dissertation was approved for publication on 2025-07-18 at 05:48.","DSpace SAF Submission Ingestion Package generated from Vireo submission #22620 on 2025-10-20 at 20:15:25","Distributed Sensor Networks (DSNs) [1] are ubiquitous in civilian and military applications, with a wide range of uses including motion detection, healthcare, remote sensing, surveillance, and imaging. The deployment of DSNs involves various constraints on the sensor nodes and the network, such as i) bandwidth-limited communication ii) power consumption iii) response time of sensor nodes iv) compute limitations and many others. For instance, a DSN deployed in a battlefield must exhibit low power consumption and fast sensor response time to external stimuli within a bandwidth-limited communication network. In recent years, computing power has undergone significant advancements, largely propelled by improvements in System-on-Chip (SoC) technologies. These improvements have paved the way for the integration of Machine Learning (ML) capabilities into sensor nodes. Machine learning algorithms often leverage neural networks, which demand substantial memory, bandwidth, and computational resources. Consequently, this imposes heightened bandwidth and power requisites on DSNs to support ML implementation in sensor nodes. Thus, there arises a pressing need for innovative algorithms aimed at resource optimization within DSNs, a challenge that we address in this thesis. Resource optimization algorithms for DSNs vary based on their topology. In this thesis, we focus on DSNs characterized by a centralized structure, comprising a central controller/central node and numerous sensor nodes. Notably, the sensors can communicate only with the central node. Within the scope of this thesis, we refer to this topology as the central topology. Within central topology DSNs, we tackle two distinct problems; i) energy efficient model-free approach to tracking an object moving through a sensing grid ii) bandwidth-efficient methods for learning a machine learning model at the central controller using data gathered from the sensor nodes. We address the problem of energy efficient model-free tracking though the lens of RL (Reinforcement Learning) and the problem of bandwidth efficient learning through compressed distributed Stochastic Gradient Descent (SGD)."]},{"key":"dc:format","label":"Dc Format","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["Reinforcement learning and optimization methods in sensor networks"]}]}],"canonical_facts":{"dc:contributor":["Veeravalli, Venugopal V.","Veeravalli, Venugopal V","Moulin, Pierre","Varshney, Lav R","Chatterjee, Sabyasachi","Zachary Hare, James"],"dc:creator":["Muthuveeru-Subramaniam, Adarsh"],"dc:date":["2025-07-18","2025-08"],"dc:description":["Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2025-10-20 without embargo terms","The student, Adarsh Muthuveeru-Subramaniam, accepted the attached license on 2025-07-16 at 19:42.","The student, Adarsh Muthuveeru-Subramaniam, submitted this Dissertation for approval on 2025-07-16 at 19:48.","This Dissertation was approved for publication on 2025-07-18 at 05:48.","DSpace SAF Submission Ingestion Package generated from Vireo submission #22620 on 2025-10-20 at 20:15:25","Distributed Sensor Networks (DSNs) [1] are ubiquitous in civilian and military applications, with a wide range of uses including motion detection, healthcare, remote sensing, surveillance, and imaging. The deployment of DSNs involves various constraints on the sensor nodes and the network, such as i) bandwidth-limited communication ii) power consumption iii) response time of sensor nodes iv) compute limitations and many others. For instance, a DSN deployed in a battlefield must exhibit low power consumption and fast sensor response time to external stimuli within a bandwidth-limited communication network. In recent years, computing power has undergone significant advancements, largely propelled by improvements in System-on-Chip (SoC) technologies. These improvements have paved the way for the integration of Machine Learning (ML) capabilities into sensor nodes. Machine learning algorithms often leverage neural networks, which demand substantial memory, bandwidth, and computational resources. Consequently, this imposes heightened bandwidth and power requisites on DSNs to support ML implementation in sensor nodes. Thus, there arises a pressing need for innovative algorithms aimed at resource optimization within DSNs, a challenge that we address in this thesis. Resource optimization algorithms for DSNs vary based on their topology. In this thesis, we focus on DSNs characterized by a centralized structure, comprising a central controller/central node and numerous sensor nodes. Notably, the sensors can communicate only with the central node. Within the scope of this thesis, we refer to this topology as the central topology. Within central topology DSNs, we tackle two distinct problems; i) energy efficient model-free approach to tracking an object moving through a sensing grid ii) bandwidth-efficient methods for learning a machine learning model at the central controller using data gathered from the sensor nodes. We address the problem of energy efficient model-free tracking though the lens of RL (Reinforcement Learning) and the problem of bandwidth efficient learning through compressed distributed Stochastic Gradient Descent (SGD)."],"dc:format":["application/pdf"],"dc:identifier":["https://hdl.handle.net/2142/129955"],"dc:language":["en","eng"],"dc:rights":["Copyright 2025 Adarsh Muthuveeru-Subramaniam"],"dc:subject":["Reinforcement Learning","Optimization","Sensor Networks"],"dc:title":["Reinforcement learning and optimization methods in sensor networks"],"dc:type":["text"],"thesis:degree_discipline":["Electrical & Computer Engr"],"thesis:degree_level":["Dissertation"],"thesis:degree_name":["Ph.D."],"thesis:institution_name":["University of Illinois Urbana-Champaign"]},"updated_at":"2026-07-22T22:25:06Z"}