{"id":{"repo_id":"uwo","oai_identifier":"oai:uwo.scholaris.ca:20.500.14721/38505"},"canonical_url":"https://search.dev.ndltd.org/etd/uwo/oai:uwo.scholaris.ca:20.500.14721/38505","repository":{"repo_id":"uwo","name":"Western University","base_url":"https://uwo.scholaris.ca/server/oai/request"},"display":{"title":"Gaze-Aware Driver Maneuver Prediction Using Object Detection and Sequential Deep Learning Models for Advanced Driver Assistance Systems","abstract":"Maneuver prediction is essential in Advanced Driver Assistance Systems (ADAS), enabling proactive interventions by anticipating driver actions based on behavioral and contextual cues. This thesis investigates maneuver prediction in ADAS by integrating driver gaze behavior, vehicle dynamics, and environmental object detection. The central objective is to evaluate whether incorporating driver attention to traffic-relevant elements, specifically, traffic signs and lights, can improve the accuracy and timeliness of maneuver predictions. Object detection was performed using custom-trained YOLOv11 models, whose outputs were aligned with gaze data to create object-in-gaze features. These features, combined with vehicle sensor data, were used to train and evaluate three deep learning architectures: Gated Recurrent Units (GRU), Long Short-Term Memory (LSTM), and the Temporal Convolutional Network (TCN). Results indicate that incorporating gaze-based object detection features slightly increased the time available to anticipate maneuvers but did not improve prediction accuracy. The Temporal Convolutional Network (TCN) model, in particular, demonstrated robust early prediction capabilities. This work highlights the value of combining behavioral and contextual cues for temporal modeling in driving scenarios and underscores the potential of TCNs for efficient and accurate maneuver prediction in future ADAS systems.","abstract_html":"Maneuver prediction is essential in Advanced Driver Assistance Systems (ADAS), enabling proactive interventions by anticipating driver actions based on behavioral and contextual cues. This thesis investigates maneuver prediction in ADAS by integrating driver gaze behavior, vehicle dynamics, and environmental object detection. The central objective is to evaluate whether incorporating driver attention to traffic-relevant elements, specifically, traffic signs and lights, can improve the accuracy and timeliness of maneuver predictions. Object detection was performed using custom-trained YOLOv11 models, whose outputs were aligned with gaze data to create object-in-gaze features. These features, combined with vehicle sensor data, were used to train and evaluate three deep learning architectures: Gated Recurrent Units (GRU), Long Short-Term Memory (LSTM), and the Temporal Convolutional Network (TCN). Results indicate that incorporating gaze-based object detection features slightly increased the time available to anticipate maneuvers but did not improve prediction accuracy. The Temporal Convolutional Network (TCN) model, in particular, demonstrated robust early prediction capabilities. This work highlights the value of combining behavioral and contextual cues for temporal modeling in driving scenarios and underscores the potential of TCNs for efficient and accurate maneuver prediction in future ADAS systems.","abstract_has_math":false,"creators":["Sawane, Ravindranath"],"institution":"The University of Western Ontario","degree_name":"M Sc","degree_level":null,"degree_discipline":"Computer Science","degree_department":null,"school":null,"contributors":[],"advisors":["Bauer, Michael"],"committee_chairs":[],"committee_members":[],"year":2025,"date_issued":"2025-08-08","date_published":"2025-08-08","updated_at":"2026-07-27T21:55:58Z","subjects":["Advanced Driver Assistance Systems (ADAS)","Driver Maneuver Predic- tion","Gaze Tracking","Vehicle Dynamics","YOLOv11","Object Detection","Temporal Modeling","Sequential Deep Learning","Temporal Convolutional Networks (TCN)","Long Short-Term Memory (LSTM)","Gated Recurrent Unit (GRU)","Sliding Window","Ottawa Driving Dataset","LISA Traffic Dataset","Eye-Tracking Data","Point of Gaze (PoG)","CANBus Signals","Early Maneuver Anticipation","Gaze-Aware Systems","Intelligent Transportation Systems"],"languages":["en"],"rights":[],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://hdl.handle.net/20.500.14721/38505","outbound_label":"Handle","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Bauer, Michael"]},{"key":"dc:creator","label":"Author","values":["Sawane, Ravindranath"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2025-08-10T20:51:22Z"]},{"key":"dc:date.available","label":"Dc Date Available","values":["2025-08-10T20:51:22Z"]},{"key":"dc:date.issued","label":"Date","values":["2025-08-08"]},{"key":"dc:publisher","label":"Institution","values":["The University of Western Ontario"]},{"key":"dc:type","label":"Dc Type","values":["thesis"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Computer Science"]},{"key":"thesis:degree_name","label":"Degree Name","values":["M Sc"]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["The University of Western Ontario"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Advanced Driver Assistance Systems (ADAS)","Driver Maneuver Predic- tion","Gaze Tracking","Vehicle Dynamics","YOLOv11","Object Detection","Temporal Modeling","Sequential Deep Learning","Temporal Convolutional Networks (TCN)","Long Short-Term Memory (LSTM)","Gated Recurrent Unit (GRU)","Sliding Window","Ottawa Driving Dataset","LISA Traffic Dataset","Eye-Tracking Data","Point of Gaze (PoG)","CANBus Signals","Early Maneuver Anticipation","Gaze-Aware Systems","Intelligent Transportation Systems"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language.iso","label":"Language (ISO)","values":["en"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://hdl.handle.net/20.500.14721/38505"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Maneuver prediction is essential in Advanced Driver Assistance Systems (ADAS), enabling proactive interventions by anticipating driver actions based on behavioral and contextual cues. This thesis investigates maneuver prediction in ADAS by integrating driver gaze behavior, vehicle dynamics, and environmental object detection. The central objective is to evaluate whether incorporating driver attention to traffic-relevant elements, specifically, traffic signs and lights, can improve the accuracy and timeliness of maneuver predictions. Object detection was performed using custom-trained YOLOv11 models, whose outputs were aligned with gaze data to create object-in-gaze features. These features, combined with vehicle sensor data, were used to train and evaluate three deep learning architectures: Gated Recurrent Units (GRU), Long Short-Term Memory (LSTM), and the Temporal Convolutional Network (TCN). Results indicate that incorporating gaze-based object detection features slightly increased the time available to anticipate maneuvers but did not improve prediction accuracy. The Temporal Convolutional Network (TCN) model, in particular, demonstrated robust early prediction capabilities. This work highlights the value of combining behavioral and contextual cues for temporal modeling in driving scenarios and underscores the potential of TCNs for efficient and accurate maneuver prediction in future ADAS systems."]},{"key":"dc:title","label":"Title","values":["Gaze-Aware Driver Maneuver Prediction Using Object Detection and Sequential Deep Learning Models for Advanced Driver Assistance Systems"]}]}],"canonical_facts":{"dc:contributor.advisor":["Bauer, Michael"],"dc:creator":["Sawane, Ravindranath"],"dc:date.accessioned":["2025-08-10T20:51:22Z"],"dc:date.available":["2025-08-10T20:51:22Z"],"dc:date.issued":["2025-08-08"],"dc:description.abstract":["Maneuver prediction is essential in Advanced Driver Assistance Systems (ADAS), enabling proactive interventions by anticipating driver actions based on behavioral and contextual cues. This thesis investigates maneuver prediction in ADAS by integrating driver gaze behavior, vehicle dynamics, and environmental object detection. The central objective is to evaluate whether incorporating driver attention to traffic-relevant elements, specifically, traffic signs and lights, can improve the accuracy and timeliness of maneuver predictions. Object detection was performed using custom-trained YOLOv11 models, whose outputs were aligned with gaze data to create object-in-gaze features. These features, combined with vehicle sensor data, were used to train and evaluate three deep learning architectures: Gated Recurrent Units (GRU), Long Short-Term Memory (LSTM), and the Temporal Convolutional Network (TCN). Results indicate that incorporating gaze-based object detection features slightly increased the time available to anticipate maneuvers but did not improve prediction accuracy. The Temporal Convolutional Network (TCN) model, in particular, demonstrated robust early prediction capabilities. This work highlights the value of combining behavioral and contextual cues for temporal modeling in driving scenarios and underscores the potential of TCNs for efficient and accurate maneuver prediction in future ADAS systems."],"dc:identifier.uri":["https://hdl.handle.net/20.500.14721/38505"],"dc:language.iso":["en"],"dc:publisher":["The University of Western Ontario"],"dc:subject":["Advanced Driver Assistance Systems (ADAS)","Driver Maneuver Predic- tion","Gaze Tracking","Vehicle Dynamics","YOLOv11","Object Detection","Temporal Modeling","Sequential Deep Learning","Temporal Convolutional Networks (TCN)","Long Short-Term Memory (LSTM)","Gated Recurrent Unit (GRU)","Sliding Window","Ottawa Driving Dataset","LISA Traffic Dataset","Eye-Tracking Data","Point of Gaze (PoG)","CANBus Signals","Early Maneuver Anticipation","Gaze-Aware Systems","Intelligent Transportation Systems"],"dc:title":["Gaze-Aware Driver Maneuver Prediction Using Object Detection and Sequential Deep Learning Models for Advanced Driver Assistance Systems"],"dc:type":["thesis"],"thesis:degree_discipline":["Computer Science"],"thesis:degree_name":["M Sc"],"thesis:institution_name":["The University of Western Ontario"]},"updated_at":"2026-07-27T21:55:58Z"}