{"id":{"repo_id":"mcmaster","oai_identifier":"oai:macsphere.mcmaster.ca:11375/33582"},"canonical_url":"https://search.ndltd.org/etd/mcmaster/oai:macsphere.mcmaster.ca:11375/33582","repository":{"repo_id":"mcmaster","name":"McMaster University","base_url":"https://macsphere.mcmaster.ca/server/oai/request"},"display":{"title":"Queue-Aware Traffic Signal Optimization for Congestion Management","abstract":"Urban traffic congestion remains a persistent challenge at signalized intersections and arterial corridors. Effective signal timing can substantially improve traffic operations, but its performance depends on the availability of reliable traffic-state information. This thesis develops a queue-aware traffic signal optimization framework that leverages information derived from conventional lane-level stopline detectors to improve congestion management without requiring connected vehicles, probe data, or extensive sensing infrastructure. The research consists of three integrated studies. The first develops a machine learning approach for identifying queued vehicles from detector data under varying traffic conditions. A random forest recovers per-vehicle queuing status with 83% accuracy and stays robust with up to 30% of detections missing. The second estimates queue length using a hybrid framework that combines deep learning and adaptive neuro-fuzzy inference techniques to provide accurate and robust queue length estimates across undersaturated, oversaturated, and incident conditions. Its mean absolute percentage error is 9.25% under oversaturation. These two studies establish the traffic-state information required for advanced signal control. Building on this foundation, the third study develops a queue-aware traffic signal optimization framework that incorporates queue information directly into the signal timing process. Estimated queues are used to dynamically adjust minimum green requirements and guide the search for feasible timing plans, improving optimization efficiency while maintaining traffic progression along corridors. The thesis therefore presents a practical and deployable framework that integrates traffic-state estimation and signal optimization on the detection infrastructure agencies already operate. Simulation experiments quantify the benefit. The queue-based constraints prune the per-intersection plan space by 62% at a volume-to-capacity ratio of 0.8. Across the demand scenarios tested, the framework reduces corridor delay relative to conventional and state-of-the-art signal control strategies. At the highest demand level evaluated, corridor delay is 72% lower than a throughput-optimal Max-pressure benchmark while throughput stays within 3.1% of it.","abstract_html":"Urban traffic congestion remains a persistent challenge at signalized intersections and arterial corridors. Effective signal timing can substantially improve traffic operations, but its performance depends on the availability of reliable traffic-state information. This thesis develops a queue-aware traffic signal optimization framework that leverages information derived from conventional lane-level stopline detectors to improve congestion management without requiring connected vehicles, probe data, or extensive sensing infrastructure. The research consists of three integrated studies. The first develops a machine learning approach for identifying queued vehicles from detector data under varying traffic conditions. A random forest recovers per-vehicle queuing status with 83% accuracy and stays robust with up to 30% of detections missing. The second estimates queue length using a hybrid framework that combines deep learning and adaptive neuro-fuzzy inference techniques to provide accurate and robust queue length estimates across undersaturated, oversaturated, and incident conditions. Its mean absolute percentage error is 9.25% under oversaturation. These two studies establish the traffic-state information required for advanced signal control. Building on this foundation, the third study develops a queue-aware traffic signal optimization framework that incorporates queue information directly into the signal timing process. Estimated queues are used to dynamically adjust minimum green requirements and guide the search for feasible timing plans, improving optimization efficiency while maintaining traffic progression along corridors. The thesis therefore presents a practical and deployable framework that integrates traffic-state estimation and signal optimization on the detection infrastructure agencies already operate. Simulation experiments quantify the benefit. The queue-based constraints prune the per-intersection plan space by 62% at a volume-to-capacity ratio of 0.8. Across the demand scenarios tested, the framework reduces corridor delay relative to conventional and state-of-the-art signal control strategies. At the highest demand level evaluated, corridor delay is 72% lower than a throughput-optimal Max-pressure benchmark while throughput stays within 3.1% of it.","abstract_has_math":false,"creators":["Gu, Yuhang"],"institution":null,"degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":"Civil Engineering","school":null,"contributors":[],"advisors":["Razavi, Saiedeh","Habibi, Saeid"],"committee_chairs":[],"committee_members":[],"year":2026,"date_issued":"2026","date_published":"2026","updated_at":"2026-11-18T19:33:44Z","subjects":["queuing status detection","queue length estimation","signal timing optimization","signalized arterials","loop detectors","ANFIS","deep learning","intelligent transportation systems"],"languages":["en"],"rights":["Attribution-NonCommercial-NoDerivatives 4.0 International"],"rights_urls":["http://creativecommons.org/licenses/by-nc-nd/4.0/"],"identifier_entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://doi.org/10.71548/7680"],"render_values":[{"text":"https://doi.org/10.71548/7680","href":"https://doi.org/10.71548/7680","code":true}]}],"person_identifier_entries":[]},"links":{"outbound_url":"https://hdl.handle.net/11375/33582","outbound_label":"Handle","outbound_source":"dc:identifier.uri","record_url":"https://hdl.handle.net/11375/33582","record_label":"Repository record","record_kind":"handle","record_source":"dc:identifier.uri","download_url":null,"download_label":null,"download_source":null,"access_status":null,"embargo_until":null,"access_note":null},"source_record":{"url":"https://macsphere.mcmaster.ca/server/oai/request?verb=GetRecord&metadataPrefix=dim&identifier=oai%3Amacsphere.mcmaster.ca%3A11375%2F33582","prefix":"dim"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Razavi, Saiedeh","Habibi, Saeid"]},{"key":"dc:contributor.department","label":"Department","values":["Civil Engineering"]},{"key":"dc:creator","label":"Author","values":["Gu, Yuhang"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2026-09-22T18:01:46Z"]},{"key":"dc:date.issued","label":"Date","values":["2026"]},{"key":"dc:type","label":"Dc Type","values":["Thesis"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["queuing status detection","queue length estimation","signal timing optimization","signalized arterials","loop detectors","ANFIS","deep learning","intelligent transportation systems"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language.iso","label":"Language (ISO)","values":["en"]},{"key":"dc:rights","label":"Dc Rights","values":["Attribution-NonCommercial-NoDerivatives 4.0 International"]},{"key":"dc:rights.uri","label":"Rights URI","values":["http://creativecommons.org/licenses/by-nc-nd/4.0/"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://hdl.handle.net/11375/33582","https://doi.org/10.71548/7680"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Urban traffic congestion remains a persistent challenge at signalized intersections and arterial corridors. Effective signal timing can substantially improve traffic operations, but its performance depends on the availability of reliable traffic-state information. This thesis develops a queue-aware traffic signal optimization framework that leverages information derived from conventional lane-level stopline detectors to improve congestion management without requiring connected vehicles, probe data, or extensive sensing infrastructure. The research consists of three integrated studies. The first develops a machine learning approach for identifying queued vehicles from detector data under varying traffic conditions. A random forest recovers per-vehicle queuing status with 83% accuracy and stays robust with up to 30% of detections missing. The second estimates queue length using a hybrid framework that combines deep learning and adaptive neuro-fuzzy inference techniques to provide accurate and robust queue length estimates across undersaturated, oversaturated, and incident conditions. Its mean absolute percentage error is 9.25% under oversaturation. These two studies establish the traffic-state information required for advanced signal control. Building on this foundation, the third study develops a queue-aware traffic signal optimization framework that incorporates queue information directly into the signal timing process. Estimated queues are used to dynamically adjust minimum green requirements and guide the search for feasible timing plans, improving optimization efficiency while maintaining traffic progression along corridors. The thesis therefore presents a practical and deployable framework that integrates traffic-state estimation and signal optimization on the detection infrastructure agencies already operate. Simulation experiments quantify the benefit. The queue-based constraints prune the per-intersection plan space by 62% at a volume-to-capacity ratio of 0.8. Across the demand scenarios tested, the framework reduces corridor delay relative to conventional and state-of-the-art signal control strategies. At the highest demand level evaluated, corridor delay is 72% lower than a throughput-optimal Max-pressure benchmark while throughput stays within 3.1% of it."]},{"key":"dc:description.degree","label":"Dc Description Degree","values":["Doctor of Philosophy (PhD)"]},{"key":"dc:title","label":"Title","values":["Queue-Aware Traffic Signal Optimization for Congestion Management"]},{"key":"others:access_status","label":"Others Access Status","values":["embargo"]}]}],"canonical_facts":{"dc:contributor.advisor":["Razavi, Saiedeh","Habibi, Saeid"],"dc:contributor.department":["Civil Engineering"],"dc:creator":["Gu, Yuhang"],"dc:date.accessioned":["2026-09-22T18:01:46Z"],"dc:date.issued":["2026"],"dc:description.abstract":["Urban traffic congestion remains a persistent challenge at signalized intersections and arterial corridors. Effective signal timing can substantially improve traffic operations, but its performance depends on the availability of reliable traffic-state information. This thesis develops a queue-aware traffic signal optimization framework that leverages information derived from conventional lane-level stopline detectors to improve congestion management without requiring connected vehicles, probe data, or extensive sensing infrastructure. The research consists of three integrated studies. The first develops a machine learning approach for identifying queued vehicles from detector data under varying traffic conditions. A random forest recovers per-vehicle queuing status with 83% accuracy and stays robust with up to 30% of detections missing. The second estimates queue length using a hybrid framework that combines deep learning and adaptive neuro-fuzzy inference techniques to provide accurate and robust queue length estimates across undersaturated, oversaturated, and incident conditions. Its mean absolute percentage error is 9.25% under oversaturation. These two studies establish the traffic-state information required for advanced signal control. Building on this foundation, the third study develops a queue-aware traffic signal optimization framework that incorporates queue information directly into the signal timing process. Estimated queues are used to dynamically adjust minimum green requirements and guide the search for feasible timing plans, improving optimization efficiency while maintaining traffic progression along corridors. The thesis therefore presents a practical and deployable framework that integrates traffic-state estimation and signal optimization on the detection infrastructure agencies already operate. Simulation experiments quantify the benefit. The queue-based constraints prune the per-intersection plan space by 62% at a volume-to-capacity ratio of 0.8. Across the demand scenarios tested, the framework reduces corridor delay relative to conventional and state-of-the-art signal control strategies. At the highest demand level evaluated, corridor delay is 72% lower than a throughput-optimal Max-pressure benchmark while throughput stays within 3.1% of it."],"dc:description.degree":["Doctor of Philosophy (PhD)"],"dc:identifier.uri":["https://hdl.handle.net/11375/33582","https://doi.org/10.71548/7680"],"dc:language.iso":["en"],"dc:rights":["Attribution-NonCommercial-NoDerivatives 4.0 International"],"dc:rights.uri":["http://creativecommons.org/licenses/by-nc-nd/4.0/"],"dc:subject":["queuing status detection","queue length estimation","signal timing optimization","signalized arterials","loop detectors","ANFIS","deep learning","intelligent transportation systems"],"dc:title":["Queue-Aware Traffic Signal Optimization for Congestion Management"],"dc:type":["Thesis"],"others:access_status":["embargo"]},"updated_at":"2026-11-18T19:33:44Z"}