{"id":{"repo_id":"houston","oai_identifier":"oai:uh-ir.tdl.org:10657/3431"},"canonical_url":"https://search.dev.ndltd.org/etd/houston/oai:uh-ir.tdl.org:10657/3431","repository":{"repo_id":"houston","name":"University of Houston","base_url":"https://uh-ir.tdl.org/server/oai/request"},"display":{"title":"Forecasting Markers of Habitual Driving Behaviors Associated with Crash Risk","abstract":"Increasingly sophisticated driver assistance systems enhance safety by issuing notifications upon sensing lane departures or applying the brakes when detecting imminent collisions. Such systems, although remarkable, are reactionary and machine centered. Here we propose a method that is mixed in its approach, preventive in its aim, and predictive in its function. The method uses multimodal measurements of the driver’s physiological variables and readings of the vehicle’s driving parameters, selecting the most informative features out of them to feed an extreme gradient boosting machine learning algorithm. The model operates upon these select features in a time window covering the recent past to make short-term predictions for the immediate future, regarding the driver’s distraction and driving style. The drivers are classified as distracted based on the presence of mental activity or physical interactions antagonistic to the driving task; their driving style is determined by steering and acceleration and is classified as aggressive or normal. Reliable short-term predictions of such behaviors, especially of repeated nature, can provide sobering awareness to the drivers, who often drift to these states subconsciously. These predictions can also inform remedial actions in future advanced driver assistance systems. The method has been tested on SIM 1 – a publicly available dataset from a major distracted driving experiment (n=59), featuring over 1.5 hours of driving for each participant and 10 channels of information, captured via unobtrusive physiological and vehicle sensors. Using 30 seconds from the immediate past, the method predicted for the next 10 seconds distracted driving with 84\\% accuracy and aggressive driving with 87\\% . It also supported previous findings and revealed further associative patterns of sympathetic arousal with driving behavior.","abstract_html":"Increasingly sophisticated driver assistance systems enhance safety by issuing notifications upon sensing lane departures or applying the brakes when detecting imminent collisions. Such systems, although remarkable, are reactionary and machine centered. Here we propose a method that is mixed in its approach, preventive in its aim, and predictive in its function. The method uses multimodal measurements of the driver’s physiological variables and readings of the vehicle’s driving parameters, selecting the most informative features out of them to feed an extreme gradient boosting machine learning algorithm. The model operates upon these select features in a time window covering the recent past to make short-term predictions for the immediate future, regarding the driver’s distraction and driving style. The drivers are classified as distracted based on the presence of mental activity or physical interactions antagonistic to the driving task; their driving style is determined by steering and acceleration and is classified as aggressive or normal. Reliable short-term predictions of such behaviors, especially of repeated nature, can provide sobering awareness to the drivers, who often drift to these states subconsciously. These predictions can also inform remedial actions in future advanced driver assistance systems. The method has been tested on SIM 1 – a publicly available dataset from a major distracted driving experiment (n=59), featuring over 1.5 hours of driving for each participant and 10 channels of information, captured via unobtrusive physiological and vehicle sensors. Using 30 seconds from the immediate past, the method predicted for the next 10 seconds distracted driving with 84\\% accuracy and aggressive driving with 87\\% . It also supported previous findings and revealed further associative patterns of sympathetic arousal with driving behavior.","abstract_has_math":false,"creators":["Panagopoulos, George 1992-"],"institution":"University of Houston","degree_name":"Master of Science","degree_level":"Masters","degree_discipline":"Computer Science","degree_department":null,"school":null,"contributors":[],"advisors":["Pavlidis, Ioannis T."],"committee_chairs":[],"committee_members":["Vilalta, Ricardo","Karkaletsis, Vangelis"],"year":2018,"date_issued":"2018-05","date_published":"2018-05","updated_at":"2026-07-24T02:33:06Z","subjects":["Affective Monitoring","Machine learning","Safety driving"],"languages":["eng"],"rights":["The author of this work is the copyright owner. UH Libraries and the Texas Digital Library have their permission to store and provide access to this work. Further transmission, reproduction, or presentation of this work is prohibited except with permission of the author(s)."],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"http://hdl.handle.net/10657/3431","outbound_label":"Handle","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Pavlidis, Ioannis T."]},{"key":"dc:contributor.committeemember","label":"Committee Member","values":["Vilalta, Ricardo","Karkaletsis, Vangelis"]},{"key":"dc:creator","label":"Author","values":["Panagopoulos, George 1992-"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2018-11-30T15:25:48Z"]},{"key":"dc:date.available","label":"Dc Date Available","values":["2018-11-30T15:25:48Z"]},{"key":"dc:date.issued","label":"Date","values":["2018-05"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Computer Science"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Masters"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Master of Science"]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["University of Houston"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Affective Monitoring","Machine learning","Safety driving"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language.iso","label":"Language (ISO)","values":["eng"]},{"key":"dc:rights","label":"Dc Rights","values":["The author of this work is the copyright owner. UH Libraries and the Texas Digital Library have their permission to store and provide access to this work. Further transmission, reproduction, or presentation of this work is prohibited except with permission of the author(s)."]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["http://hdl.handle.net/10657/3431"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Increasingly sophisticated driver assistance systems enhance safety by issuing notifications upon sensing lane departures or applying the brakes when detecting imminent collisions. Such systems, although remarkable, are reactionary and machine centered. Here we propose a method that is mixed in its approach, preventive in its aim, and predictive in its function. The method uses multimodal measurements of the driver’s physiological variables and readings of the vehicle’s driving parameters, selecting the most informative features out of them to feed an extreme gradient boosting machine learning algorithm. The model operates upon these select features in a time window covering the recent past to make short-term predictions for the immediate future, regarding the driver’s distraction and driving style. The drivers are classified as distracted based on the presence of mental activity or physical interactions antagonistic to the driving task; their driving style is determined by steering and acceleration and is classified as aggressive or normal. Reliable short-term predictions of such behaviors, especially of repeated nature, can provide sobering awareness to the drivers, who often drift to these states subconsciously. These predictions can also inform remedial actions in future advanced driver assistance systems. The method has been tested on SIM 1 – a publicly available dataset from a major distracted driving experiment (n=59), featuring over 1.5 hours of driving for each participant and 10 channels of information, captured via unobtrusive physiological and vehicle sensors. Using 30 seconds from the immediate past, the method predicted for the next 10 seconds distracted driving with 84\\% accuracy and aggressive driving with 87\\% . It also supported previous findings and revealed further associative patterns of sympathetic arousal with driving behavior."]},{"key":"dc:format.mimetype","label":"Dc Format Mimetype","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["Forecasting Markers of Habitual Driving Behaviors Associated with Crash Risk"]}]}],"canonical_facts":{"dc:contributor.advisor":["Pavlidis, Ioannis T."],"dc:contributor.committeemember":["Vilalta, Ricardo","Karkaletsis, Vangelis"],"dc:creator":["Panagopoulos, George 1992-"],"dc:date.accessioned":["2018-11-30T15:25:48Z"],"dc:date.available":["2018-11-30T15:25:48Z"],"dc:date.issued":["2018-05"],"dc:description.abstract":["Increasingly sophisticated driver assistance systems enhance safety by issuing notifications upon sensing lane departures or applying the brakes when detecting imminent collisions. Such systems, although remarkable, are reactionary and machine centered. Here we propose a method that is mixed in its approach, preventive in its aim, and predictive in its function. The method uses multimodal measurements of the driver’s physiological variables and readings of the vehicle’s driving parameters, selecting the most informative features out of them to feed an extreme gradient boosting machine learning algorithm. The model operates upon these select features in a time window covering the recent past to make short-term predictions for the immediate future, regarding the driver’s distraction and driving style. The drivers are classified as distracted based on the presence of mental activity or physical interactions antagonistic to the driving task; their driving style is determined by steering and acceleration and is classified as aggressive or normal. Reliable short-term predictions of such behaviors, especially of repeated nature, can provide sobering awareness to the drivers, who often drift to these states subconsciously. These predictions can also inform remedial actions in future advanced driver assistance systems. The method has been tested on SIM 1 – a publicly available dataset from a major distracted driving experiment (n=59), featuring over 1.5 hours of driving for each participant and 10 channels of information, captured via unobtrusive physiological and vehicle sensors. Using 30 seconds from the immediate past, the method predicted for the next 10 seconds distracted driving with 84\\% accuracy and aggressive driving with 87\\% . It also supported previous findings and revealed further associative patterns of sympathetic arousal with driving behavior."],"dc:format.mimetype":["application/pdf"],"dc:identifier.uri":["http://hdl.handle.net/10657/3431"],"dc:language.iso":["eng"],"dc:rights":["The author of this work is the copyright owner. UH Libraries and the Texas Digital Library have their permission to store and provide access to this work. Further transmission, reproduction, or presentation of this work is prohibited except with permission of the author(s)."],"dc:subject":["Affective Monitoring","Machine learning","Safety driving"],"dc:title":["Forecasting Markers of Habitual Driving Behaviors Associated with Crash Risk"],"thesis:degree_discipline":["Computer Science"],"thesis:degree_level":["Masters"],"thesis:degree_name":["Master of Science"],"thesis:institution_name":["University of Houston"]},"updated_at":"2026-07-24T02:33:06Z"}