Shan Bao
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Shan Bao has a total of 113 co-authors. Here are the top co-authors by number of shared publications:
- Bo Yu
Key Laboratory of Road and Traffic Engineering of the Ministry of Education, College of Transportation Engineering, Tongji University, Shanghai 201804, China
7 shared publications - Huizhong Guo
University of Michigan Transportation Research Institute, Ann Arbor, MI
7 shared publications - Yuren Chen
Key Laboratory of Road and Traffic Engineering of the Ministry of Education, College of Transportation Engineering, Tongji University, 4800 Cao'an Highway, Shanghai, 201804, China
7 shared publications - David J. LeBlanc
University of Michigan Transportation Research Institute, Ann Arbor, MI, USA
5 shared publications Huizhong Guo
University of Michigan Transportation Research Institute
5 shared publicationsBo Yu
Key Laboratory of Road and Traffic Engineering of the Ministry of Education, College of Transportation Engineering, Tongji University, Shanghai, China
4 shared publications
Latest Publications
- Open Access
- 20 October 2026
- Published by Taylor & Francis in IISE Transactions on Occupational Ergonomics and Human Factors
- p. 1-21
Trust, Traits, and Turning Points: Decoding Operator Responses to Real-World ADAS Malfunctions
1. As vehicles adopt increasingly sophisticated automation systems, driving is becoming more of a managerial task rather than a manual one. The shift from manual control to supervisory system manag...
- 20 October 2026
- Published by Taylor & Francis in IISE Transactions on Occupational Ergonomics and Human Factors
- p. 1-21
Open AccessTrust, Traits, and Turning Points: Decoding Operator Responses to Real-World ADAS Malfunctions
1. As vehicles adopt increasingly sophisticated automation systems, driving is becoming more of a managerial task rather than a manual one. The shift from manual control to supervisory system manag...
- Open Access
- 7 July 2026
- Published by Elsevier in SSRN Electronic Journal
Artificial Intelligence for Road User Behavior: A Scoping Review of Current Paradigms and Future Frontiers
Road user behavior remains the primary precipitating factor in global traffic crashes, yet the methodologies used to analyze these complex human dynamics are undergoing a fundamental paradigm shift. Traditional statistical inference, once the cornerstone of safety research, is increasingly being superseded by high-dimensional Artificial Intelligence (AI) modeling capable of capturing non-linear relationships. This scoping review synthesizes 140 peer-reviewed studies published between 2019 and 2026 to systematically map the state-of-the-art AI-driven behavioral analysis within crash-related contexts. To structure this evolving landscape, we introduce an original six-layer analytical taxonomy. This framework traces the intricate pathways connecting behavioral categories (such as impairment, distraction, and tactical maneuvers) to computational domains (including Machine Learning, Deep Learning, and Natural Language Processing), functional objectives, and evaluation indicators. Through Sankey-based operationalization, we analyze the research trajectories for both drivers (128 studies) and vulnerable road users (52 studies). Our synthesis reveals a field currently dominated by ”predictive fidelity”, optimizing model accuracy, often at the significant expense of ”operational viability” and real-time safety metrics required for field deployment. Notably, the recent surge in Large Language Models (LLMs) across 31 studies marks a pivotal transition toward semantic reasoning, allowing researchers to move beyond numerical data toward narrative-based information extraction and automated scenario generation. By highlighting critical structural gaps in explainability, temporal relevance, and real-world implementation, this review provides a strategic roadmap for the next generation of AI-integrated traffic safety research, advocating for a transition from black-box prediction to deployable, human-centric safety intelligence.
- 7 July 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessArtificial Intelligence for Road User Behavior: A Scoping Review of Current Paradigms and Future Frontiers
Road user behavior remains the primary precipitating factor in global traffic crashes, yet the methodologies used to analyze these complex human dynamics are undergoing a fundamental paradigm shift. Traditional statistical inference, once the cornerstone of safety research, is increasingly being superseded by high-dimensional Artificial Intelligence (AI) modeling capable of capturing non-linear relationships. This scoping review synthesizes 140 peer-reviewed studies published between 2019 and 2026 to systematically map the state-of-the-art AI-driven behavioral analysis within crash-related contexts. To structure this evolving landscape, we introduce an original six-layer analytical taxonomy. This framework traces the intricate pathways connecting behavioral categories (such as impairment, distraction, and tactical maneuvers) to computational domains (including Machine Learning, Deep Learning, and Natural Language Processing), functional objectives, and evaluation indicators. Through Sankey-based operationalization, we analyze the research trajectories for both drivers (128 studies) and vulnerable road users (52 studies). Our synthesis reveals a field currently dominated by ”predictive fidelity”, optimizing model accuracy, often at the significant expense of ”operational viability” and real-time safety metrics required for field deployment. Notably, the recent surge in Large Language Models (LLMs) across 31 studies marks a pivotal transition toward semantic reasoning, allowing researchers to move beyond numerical data toward narrative-based information extraction and automated scenario generation. By highlighting critical structural gaps in explainability, temporal relevance, and real-world implementation, this review provides a strategic roadmap for the next generation of AI-integrated traffic safety research, advocating for a transition from black-box prediction to deployable, human-centric safety intelligence.
- Open Access
- 2 July 2026
- Published by Taylor & Francis in Traffic Injury Prevention
- Vol. 27 (sup1) , 1-12
Seeing the automation malfunction before it happens: designing information cues that support driver situation awareness
Advanced Driver Assistance Systems (ADAS) may degrade in complex environments, producing malfunction states that require driver readiness to intervene. However, empirical guidance remains limited on how vehicle-driver communication should be structured to support situation awareness (SA) during such events. This study evaluates three structured information cue functions, Compromised Function (CF), Tracking Status (TS), and Action Rationale (AR), designed to support perception (Occurrence), comprehension (Cause), and projection (Solution) during ADAS malfunctions. Twenty-seven ADAS-equipped drivers evaluated cue effectiveness across three representative malfunction scenarios using a mixed experimental design. Linear mixed-effects models assessed absolute effectiveness, relative differences among cue types, and demographic moderation. CF demonstrated the most stable and robust advantage at lower SA levels. At the Occurrence stage, CF was rated significantly higher than TS (Δ = 1.33) and AR (Δ = 0.93). A similar pattern emerged at the Cause stage, where CF exceeded AR (Δ = 0.73) and TS (Δ = 0.72). Demographic moderation effects, including age, gender, education, and ADAS experience, were consistently observed, indicating stable differences in cue effectiveness across user groups. These findings highlight the importance of clearly communicating system vulnerabilities to support early malfunction recognition and to guide more effective transparency design for drivers facing automation malfunctions.
- 2 July 2026
- Published by Taylor & Francis in Traffic Injury Prevention
- Vol. 27 (sup1) , 1-12
Open AccessSeeing the automation malfunction before it happens: designing information cues that support driver situation awareness
Advanced Driver Assistance Systems (ADAS) may degrade in complex environments, producing malfunction states that require driver readiness to intervene. However, empirical guidance remains limited on how vehicle-driver communication should be structured to support situation awareness (SA) during such events. This study evaluates three structured information cue functions, Compromised Function (CF), Tracking Status (TS), and Action Rationale (AR), designed to support perception (Occurrence), comprehension (Cause), and projection (Solution) during ADAS malfunctions. Twenty-seven ADAS-equipped drivers evaluated cue effectiveness across three representative malfunction scenarios using a mixed experimental design. Linear mixed-effects models assessed absolute effectiveness, relative differences among cue types, and demographic moderation. CF demonstrated the most stable and robust advantage at lower SA levels. At the Occurrence stage, CF was rated significantly higher than TS (Δ = 1.33) and AR (Δ = 0.93). A similar pattern emerged at the Cause stage, where CF exceeded AR (Δ = 0.73) and TS (Δ = 0.72). Demographic moderation effects, including age, gender, education, and ADAS experience, were consistently observed, indicating stable differences in cue effectiveness across user groups. These findings highlight the importance of clearly communicating system vulnerabilities to support early malfunction recognition and to guide more effective transparency design for drivers facing automation malfunctions.
- 28 June 2026
- Published by Elsevier in Accident Analysis & Prevention
- Vol. 235, 108642
The state-dependent causal effect of a V2X-based beyond-line-of-sight warning system on young driver response: a causal machine learning approach
- 28 June 2026
- Published by Elsevier in Accident Analysis & Prevention
- Vol. 235, 108642
The state-dependent causal effect of a V2X-based beyond-line-of-sight warning system on young driver response: a causal machine learning approach
- Open Access
- 12 June 2026
- Published by Elsevier in Transportation Research Interdisciplinary Perspectives
- Vol. 38
Enhancing transportation education to reflect technological advancements in connected and automated vehicles
- 12 June 2026
- Published by Elsevier in Transportation Research Interdisciplinary Perspectives
- Vol. 38
Open AccessEnhancing transportation education to reflect technological advancements in connected and automated vehicles
- Open Access
- 18 May 2026
- Published by Elsevier in Artificial Intelligence for Transportation
- Vol. 7
From narratives to probabilistic quantification: Predicting and interpreting drivers’ hazardous actions in crashes using large language model
- 18 May 2026
- Published by Elsevier in Artificial Intelligence for Transportation
- Vol. 7
Open AccessFrom narratives to probabilistic quantification: Predicting and interpreting drivers’ hazardous actions in crashes using large language model
- Open Access
- 15 May 2026
- Published by Taylor & Francis in Traffic Injury Prevention
- p. 1-10
An examination of real-world disengagement patterns of automated driving systems in autonomous-mode: A deep-learning method
Automated Driving Systems (ADS) are rapidly advancing toward higher levels of autonomy, surpassing traditional Advanced Driver Assistance Systems (ADAS) to enable conditional automation. A key chal...
- 15 May 2026
- Published by Taylor & Francis in Traffic Injury Prevention
- p. 1-10
Open AccessAn examination of real-world disengagement patterns of automated driving systems in autonomous-mode: A deep-learning method
Automated Driving Systems (ADS) are rapidly advancing toward higher levels of autonomy, surpassing traditional Advanced Driver Assistance Systems (ADAS) to enable conditional automation. A key chal...
- Open Access
- 3 March 2026
- Published by Elsevier in SSRN Electronic Journal
Modeling Driver Compliance with Eco-Routing Recommendations Using Naturalistic Smartphone Data
Eco-routing navigation systems represent essential technology for reducing fuel consumption and emissions, yet limited research has examined whether and why drivers adopt and comply with eco-routing recommendations. This study developed a smartphone-based eco-routing navigation app and collected naturalistic driving data to analyze drivers’ route choice and compliance behavior using mixed models and an ensemble of classifier chains (ECC) model, based on driver characteristics, subjective data, and route information. Results showed that drivers were more likely to choose eco-routes when the additional time cost compared to the fastest route was smaller and the trip was shorter and had higher per-mile fuel consumption. On average, participants complied with the recommended route 56.7% of the time. Additionally, the accuracy and Hamming loss of the ECC model for route-choice behavior prediction were 92.6% and 0.026, respectively. These findings can inform effective eco-routing navigation systems design, supporting the development of more eco-friendly transportation systems.
- 3 March 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessModeling Driver Compliance with Eco-Routing Recommendations Using Naturalistic Smartphone Data
Eco-routing navigation systems represent essential technology for reducing fuel consumption and emissions, yet limited research has examined whether and why drivers adopt and comply with eco-routing recommendations. This study developed a smartphone-based eco-routing navigation app and collected naturalistic driving data to analyze drivers’ route choice and compliance behavior using mixed models and an ensemble of classifier chains (ECC) model, based on driver characteristics, subjective data, and route information. Results showed that drivers were more likely to choose eco-routes when the additional time cost compared to the fastest route was smaller and the trip was shorter and had higher per-mile fuel consumption. On average, participants complied with the recommended route 56.7% of the time. Additionally, the accuracy and Hamming loss of the ECC model for route-choice behavior prediction were 92.6% and 0.026, respectively. These findings can inform effective eco-routing navigation systems design, supporting the development of more eco-friendly transportation systems.
- 11 October 2025
- Published by Elsevier in Journal of Safety Research
- Vol. 95, 197-210
Assessing the effectiveness of driver training interventions in improving safe engagement with vehicle automation systems
- 11 October 2025
- Published by Elsevier in Journal of Safety Research
- Vol. 95, 197-210
Assessing the effectiveness of driver training interventions in improving safe engagement with vehicle automation systems
- 6 September 2025
- Published by Elsevier in Accident Analysis & Prevention
- Vol. 222, 108223
Investigating pedestrian crash injury patterns: A comparative study of children and non-children
- 6 September 2025
- Published by Elsevier in Accident Analysis & Prevention
- Vol. 222, 108223
Investigating pedestrian crash injury patterns: A comparative study of children and non-children