Ku-Lin Wen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Ku-Lin Wen has a total of 15 co-authors. Here are the top co-authors by number of shared publications:
- Tien-Pen Hsu
Department of Civil Engineering, National Taiwan University, Taipei 106, Taiwan
3 shared publications - Wei-Hsun Lee
National Cheng Kung University
3 shared publications Chi-Yun Teng
1 shared publicationChiang Fu
1 shared publication- Chih-Liang Hsiao
Department of Transportation and Communication Management Science, National Cheng Kung University, Tainan, Taiwan
1 shared publication Chin‐Hung Liu
Department of Civil Engineering National Taiwan University Taipei 106 Taiwan
1 shared publication
Latest Publications
- Open Access
- 14 July 2026
- Published by Elsevier in SSRN Electronic Journal
Driver Attention and Behavioral Anomalies Among Older and Younger Drivers During Navigation-Assisted Driving: Evidence from Real-World Eye-Tracking and Vehicle-Dynamics Data
This study examines how navigation-assisted driving influences visual attention and behavioral anomalies, with particular emphasis on age-related differences between older and younger drivers. Real-world driving data were collected under navigation-assisted and non-navigation conditions using Tobii Pro Glasses 3 and the Physics Toolbox Sensor Suite. Fourteen areas of interest (AOIs) were initially defined at the object level, including the navigation interface, pedestrians, forward roadway, right-side window, left-side window, interior rearview mirror, left-side mirror, right-side mirror, instrument cluster, motorcycles and bicycles, vehicles with four or more wheels, traffic signs, traffic signals, and other roadside elements. For statistical analysis and interpretation, these 14 AOIs were aggregated into seven functional categories: navigation interface, forward roadway, other road users, in-vehicle facilities, left-side in-vehicle facilities, signs and signals, and others. Statistical analyses examined the effects of navigation assistance and driver age on fixation behavior, saccadic activity, and AOI-specific attention allocation. An autoencoder (AE) was also implemented to identify behavioral anomalies, using reconstruction error as a surrogate anomaly indicator and the 95th percentile of the training-set error distribution as the alert threshold. Navigation assistance significantly reduced fixation counts for other road users, suggesting a redistribution of visual attention away from external dynamic hazards. Under navigation-assisted driving, older drivers showed significantly more fixations and saccades toward signs and signals, while their higher attention to the navigation interface was observed as a non-significant tendency. AE-based alerts were descriptively more frequent among older drivers and slightly more frequent under navigation-assisted driving, indicating possible reductions in behavioral consistency rather than direct evidence of crash risk. These findings provide implications for older-driver safety assessment, navigation-interface design, and advanced driver assistance systems.
- 14 July 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessDriver Attention and Behavioral Anomalies Among Older and Younger Drivers During Navigation-Assisted Driving: Evidence from Real-World Eye-Tracking and Vehicle-Dynamics Data
This study examines how navigation-assisted driving influences visual attention and behavioral anomalies, with particular emphasis on age-related differences between older and younger drivers. Real-world driving data were collected under navigation-assisted and non-navigation conditions using Tobii Pro Glasses 3 and the Physics Toolbox Sensor Suite. Fourteen areas of interest (AOIs) were initially defined at the object level, including the navigation interface, pedestrians, forward roadway, right-side window, left-side window, interior rearview mirror, left-side mirror, right-side mirror, instrument cluster, motorcycles and bicycles, vehicles with four or more wheels, traffic signs, traffic signals, and other roadside elements. For statistical analysis and interpretation, these 14 AOIs were aggregated into seven functional categories: navigation interface, forward roadway, other road users, in-vehicle facilities, left-side in-vehicle facilities, signs and signals, and others. Statistical analyses examined the effects of navigation assistance and driver age on fixation behavior, saccadic activity, and AOI-specific attention allocation. An autoencoder (AE) was also implemented to identify behavioral anomalies, using reconstruction error as a surrogate anomaly indicator and the 95th percentile of the training-set error distribution as the alert threshold. Navigation assistance significantly reduced fixation counts for other road users, suggesting a redistribution of visual attention away from external dynamic hazards. Under navigation-assisted driving, older drivers showed significantly more fixations and saccades toward signs and signals, while their higher attention to the navigation interface was observed as a non-significant tendency. AE-based alerts were descriptively more frequent among older drivers and slightly more frequent under navigation-assisted driving, indicating possible reductions in behavioral consistency rather than direct evidence of crash risk. These findings provide implications for older-driver safety assessment, navigation-interface design, and advanced driver assistance systems.
- Open Access
- 14 July 2026
- Published by Elsevier in SSRN Electronic Journal
Time Pressure, Distraction, and Vulnerability in Pedestrian Crossing Patterns at Signalized Intersections: A Latent Class Mixed Logit Analysis
Pedestrian noncompliance at signalized intersections reflects diverse responses to time pressure, distraction, vulnerability, and context, not just rule violation. This study investigates behavioral heterogeneity in pedestrian crossing patterns at signalized intersections. Using multi-camera video observations of 7,215 pedestrian crossings at six intersections in Tainan, Taiwan, each crossing event was coded as one of five mutually exclusive patterns: jaywalking between the crosswalk and the stop line, jaywalking between the crosswalk and the curb extension, partial crosswalk use, incomplete crosswalk use, and full crosswalk use. A latent class mixed multinomial logit model was applied to examine how crosswalk design, pedestrian characteristics, distraction, signal violations, walking speed, and crossing time shape these patterns. The results identify two behaviorally interpretable groups: time-pressured pedestrians and vulnerable/risk-adjusting pedestrians. Time-pressured pedestrians are more strongly associated with signal timing constraints, distraction, running behavior, jaywalking, and incomplete crosswalk use. Vulnerable and risk-adjusting pedestrians, including more older adults, mobility-impaired pedestrians, and slower walkers, generally exhibit more cautious crossing behavior but may still deviate from the crosswalk because of physical or spatial constraints. Elasticity results show that failing to complete a crossing before the all-red phase increases the likelihood of jaywalking near the curb extension, while distraction-related behaviors increase the use of partial or incomplete crosswalks. These findings demonstrate that pedestrian crossing behavior is shaped by heterogeneous behavioral responses rather than uniform compliance tendencies. The results support behaviorally informed safety strategies that distinguish urgency-driven shortcutting, distraction-related path instability, and vulnerability-related risk adjustment, rather than relying on uniform compliance-based pedestrian management.
- 14 July 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessTime Pressure, Distraction, and Vulnerability in Pedestrian Crossing Patterns at Signalized Intersections: A Latent Class Mixed Logit Analysis
Pedestrian noncompliance at signalized intersections reflects diverse responses to time pressure, distraction, vulnerability, and context, not just rule violation. This study investigates behavioral heterogeneity in pedestrian crossing patterns at signalized intersections. Using multi-camera video observations of 7,215 pedestrian crossings at six intersections in Tainan, Taiwan, each crossing event was coded as one of five mutually exclusive patterns: jaywalking between the crosswalk and the stop line, jaywalking between the crosswalk and the curb extension, partial crosswalk use, incomplete crosswalk use, and full crosswalk use. A latent class mixed multinomial logit model was applied to examine how crosswalk design, pedestrian characteristics, distraction, signal violations, walking speed, and crossing time shape these patterns. The results identify two behaviorally interpretable groups: time-pressured pedestrians and vulnerable/risk-adjusting pedestrians. Time-pressured pedestrians are more strongly associated with signal timing constraints, distraction, running behavior, jaywalking, and incomplete crosswalk use. Vulnerable and risk-adjusting pedestrians, including more older adults, mobility-impaired pedestrians, and slower walkers, generally exhibit more cautious crossing behavior but may still deviate from the crosswalk because of physical or spatial constraints. Elasticity results show that failing to complete a crossing before the all-red phase increases the likelihood of jaywalking near the curb extension, while distraction-related behaviors increase the use of partial or incomplete crosswalks. These findings demonstrate that pedestrian crossing behavior is shaped by heterogeneous behavioral responses rather than uniform compliance tendencies. The results support behaviorally informed safety strategies that distinguish urgency-driven shortcutting, distraction-related path instability, and vulnerability-related risk adjustment, rather than relying on uniform compliance-based pedestrian management.
- 6 December 2025
- Published by Springer Nature in Information Systems Frontiers
- p. 1-13
Promoting Sustainability in Motor Insurance: A Novel Approach To Evaluating Driving Behavior
- 6 December 2025
- Published by Springer Nature in Information Systems Frontiers
- p. 1-13
Promoting Sustainability in Motor Insurance: A Novel Approach To Evaluating Driving Behavior
- 13 March 2025
- Published by Taylor & Francis in Journal of Intelligent Transportation Systems
- Vol. 30 (3) , 1-16
Risky scooter riders pay more: a usage-based insurance study of a scooter risk-aware fingerprint model without privacy issues
- 13 March 2025
- Published by Taylor & Francis in Journal of Intelligent Transportation Systems
- Vol. 30 (3) , 1-16
Risky scooter riders pay more: a usage-based insurance study of a scooter risk-aware fingerprint model without privacy issues
- 1 March 2025
- Published by American Society of Civil Engineers (ASCE) in Journal of Transportation Engineering, Part A: Systems
- Vol. 151 (3)
Effect Assessment of a Motorcycle Collision Warning System at Intersections
- 1 March 2025
- Published by American Society of Civil Engineers (ASCE) in Journal of Transportation Engineering, Part A: Systems
- Vol. 151 (3)
Effect Assessment of a Motorcycle Collision Warning System at Intersections
- 20 December 2021
- Published by Taylor & Francis in Traffic Injury Prevention
- Vol. 23 (1) , 46-50
Using multinomial regression to explore the spatial factors affecting left-turn oncoming accidents involving motorcycles
In mixed traffic with a high proportion of motorcycles, accidents involving motorcycles usually account for the highest percentage of total accidents and are thus important. Because the motorcycle ...
- 20 December 2021
- Published by Taylor & Francis in Traffic Injury Prevention
- Vol. 23 (1) , 46-50
Using multinomial regression to explore the spatial factors affecting left-turn oncoming accidents involving motorcycles
In mixed traffic with a high proportion of motorcycles, accidents involving motorcycles usually account for the highest percentage of total accidents and are thus important. Because the motorcycle ...
- Open Access
- 13 October 2021
- Published by Institution of Engineering and Technology (IET) in IET Intelligent Transport Systems
- Vol. 16 (1) , 13-23
Safety effect analysis of motorcycle V2I collision warning system
Connected vehicle technology can help the driver make decisions, improve awareness of the road environment, and thereby possibly enhance traffic safety. This study explored the impact of a motorcycle vehicle‐to‐infrastructure (V2I) warning system on road safety through a field trial of a Motorcycle Safety Warning System (MSWS). This study used principal component analysis, K‐means, and ordinal logistic regression to explore the effects of the MSWS. After system activation, the road safety level (RSL) improved at 50% of the field trial sites. This study provides a reference for evaluating RSLs through the composite index of this study and verifies that the V2I warning signs of the MSWS provide a certain degree of improvement of motorcycle traffic safety. A V2I system like the MSWS can be used to collect data on driving behaviours in a short period for safety effect analysis, as compared to crash data, which require an extended period. In addition, the traffic behaviour data collected by a V2I system can also be used to identify potential high crash risk locations. The warning signs developed in this study can prevent accidents by alerting drivers who are speeding and at risk of crashes.
- 13 October 2021
- Published by Institution of Engineering and Technology (IET) in IET Intelligent Transport Systems
- Vol. 16 (1) , 13-23
Open AccessSafety effect analysis of motorcycle V2I collision warning system
Connected vehicle technology can help the driver make decisions, improve awareness of the road environment, and thereby possibly enhance traffic safety. This study explored the impact of a motorcycle vehicle‐to‐infrastructure (V2I) warning system on road safety through a field trial of a Motorcycle Safety Warning System (MSWS). This study used principal component analysis, K‐means, and ordinal logistic regression to explore the effects of the MSWS. After system activation, the road safety level (RSL) improved at 50% of the field trial sites. This study provides a reference for evaluating RSLs through the composite index of this study and verifies that the V2I warning signs of the MSWS provide a certain degree of improvement of motorcycle traffic safety. A V2I system like the MSWS can be used to collect data on driving behaviours in a short period for safety effect analysis, as compared to crash data, which require an extended period. In addition, the traffic behaviour data collected by a V2I system can also be used to identify potential high crash risk locations. The warning signs developed in this study can prevent accidents by alerting drivers who are speeding and at risk of crashes.
- Open Access
- 8 July 2021
- Published by MDPI AG in Sustainability
- Vol. 13 (14) , 7656
Applying Machine Learning to Develop Lane Control Principles for Mixed Traffic
The mixed traffic environment often has high accident rates. Therefore, many motorcycle-related traffic improvements or control methods are employed in countries with mixed traffic, including slow-traffic lanes, motorcycle two-stage left turn areas, and motorcycle waiting zones. In Taiwan, motorcycles can ride in only the two outermost lanes, including the curb lane and a mixed traffic lane. This study analyzed the new motorcycle-riding space control policy on 27 major arterial roads containing 248 road segments in Taipei by analyzing before-and-after accident data from the years 2012–2018. In this study, the equivalent-property-damage-only (EPDO) method was used to evaluate the severity of crashes before and after the cancelation of the third lane prohibition of motorcycles (TLPM) policy. After EPDO analysis, the random forest analysis method was used to screen the crucial factors in accidents for specific road segments. Finally, a classification and regression tree (CART) was created to predict the accident improvement effects of the road segments with discontinued TLPM in different situations. Furthermore, to provide practical applications, this study integrated the CART results and the needs of traffic authorities to determine four rules for canceling TLPM. In the future, on the accident-prone road segment with TLPM, the inspection of the four rules can provide the authority to decide whether to cancel TLPM to improve the accident or not.
- 8 July 2021
- Published by MDPI AG in Sustainability
- Vol. 13 (14) , 7656
Open AccessApplying Machine Learning to Develop Lane Control Principles for Mixed Traffic
The mixed traffic environment often has high accident rates. Therefore, many motorcycle-related traffic improvements or control methods are employed in countries with mixed traffic, including slow-traffic lanes, motorcycle two-stage left turn areas, and motorcycle waiting zones. In Taiwan, motorcycles can ride in only the two outermost lanes, including the curb lane and a mixed traffic lane. This study analyzed the new motorcycle-riding space control policy on 27 major arterial roads containing 248 road segments in Taipei by analyzing before-and-after accident data from the years 2012–2018. In this study, the equivalent-property-damage-only (EPDO) method was used to evaluate the severity of crashes before and after the cancelation of the third lane prohibition of motorcycles (TLPM) policy. After EPDO analysis, the random forest analysis method was used to screen the crucial factors in accidents for specific road segments. Finally, a classification and regression tree (CART) was created to predict the accident improvement effects of the road segments with discontinued TLPM in different situations. Furthermore, to provide practical applications, this study integrated the CART results and the needs of traffic authorities to determine four rules for canceling TLPM. In the future, on the accident-prone road segment with TLPM, the inspection of the four rules can provide the authority to decide whether to cancel TLPM to improve the accident or not.
- 1 June 2019
- Published by Elsevier in Journal of Safety Research
- Vol. 69, 167-176
Effect of novel divergence markings on conflict prevention regarding motorcycle-involved right turn accidents of mixed traffic flow
- 1 June 2019
- Published by Elsevier in Journal of Safety Research
- Vol. 69, 167-176
Effect of novel divergence markings on conflict prevention regarding motorcycle-involved right turn accidents of mixed traffic flow