Microscopic behaviour analysis using video recordings : A perspective on vulnerable road users
Author
Summary, in English
This thesis investigates the application of microscopic traffic data derived from video recordings to analyse the behaviour of road users, specifically focusing on pedestrians and cyclists in urban traffic environments. This research addresses the critical need for more accurate and proactive methods in the fields of traffic safety, transport system operational performance, and level of service assessment. The overall aim is to explore the potential of video technology to capture dynamic, real-time behaviours that are otherwise difficult to observe, facilitating a deeper understanding of road users’ micro-level interaction with and responses to various traffic conditions. A significant portion of the thesis is dedicated to traffic safety, using conflict-based safety analysis and exploring the integration of Surrogate Measures of Safety (SMoSs) with the Safe System approach to shift the focus from reducing ‘collisions’ to eliminating ‘severe injuries’.
The thesis includes four research papers, each contributing to the study’s overall aims and objectives. The first two papers explore methodologies for investigating the impact of physical environmental features, such as facility surfacing and outdoor lighting, on cyclist and pedestrian behaviour. The results show that the proposed method effectively captures behavioural responses to infrastructure changes, revealing that cyclists and pedestrians exhibit distinct reactions. This finding highlights the importance of treating these groups separately to address their unique needs during urban planning and infrastructure improvement processes.
The final two papers address gaps in existing safety frameworks. One proposes a refined, proactive approach to traffic safety assessment and improvement, while the other examines the severity of non-collision traffic events. By analysing video-recorded traffic situations, the study—also forming the final paper included in this thesis—identifies key objective indicators that align with human perceptions of traffic danger, emphasising the importance of two variables—proximity and potential collision consequences—in severity assessments and suggesting their integration into safety evaluations.
The studies included in this thesis demonstrate that video recordings can not only capture simple variables (e.g. the number of road users) but also provide detailed insights into road user behaviours at a micro level. By offering a clear and continuous view of traffic conditions, video data can significantly enhance the ‘intelligence input’ into traffic analysis, transforming traffic videos into actionable insights.
The thesis includes four research papers, each contributing to the study’s overall aims and objectives. The first two papers explore methodologies for investigating the impact of physical environmental features, such as facility surfacing and outdoor lighting, on cyclist and pedestrian behaviour. The results show that the proposed method effectively captures behavioural responses to infrastructure changes, revealing that cyclists and pedestrians exhibit distinct reactions. This finding highlights the importance of treating these groups separately to address their unique needs during urban planning and infrastructure improvement processes.
The final two papers address gaps in existing safety frameworks. One proposes a refined, proactive approach to traffic safety assessment and improvement, while the other examines the severity of non-collision traffic events. By analysing video-recorded traffic situations, the study—also forming the final paper included in this thesis—identifies key objective indicators that align with human perceptions of traffic danger, emphasising the importance of two variables—proximity and potential collision consequences—in severity assessments and suggesting their integration into safety evaluations.
The studies included in this thesis demonstrate that video recordings can not only capture simple variables (e.g. the number of road users) but also provide detailed insights into road user behaviours at a micro level. By offering a clear and continuous view of traffic conditions, video data can significantly enhance the ‘intelligence input’ into traffic analysis, transforming traffic videos into actionable insights.
Department/s
Publishing year
2025-03-25
Language
English
Publication/Series
Bulletin
Issue
337
Full text
Document type
Dissertation
Publisher
Department of Technology and Society, Lund University
Topic
- Transport Systems and Logistics
Keywords
- behaviour
- conflict-based analysis
- cyclist
- interactions
- microscopic traffic data
- outdoor lighting
- pedestrian
- Safe System
- Surrogate Measure of Safety
- surfacing
- traffic event severity
- traffic safety
- video analysis
Status
Published
Project
- The Third Eye
- Belysa—the impact of environmental factors on pedestrian and cyclist behavior at dusk and in the dark
Research group
- Traffic Safety and Behaviour
ISBN/ISSN/Other
- ISSN: 1653-1930
- ISSN: 1653-1930
- ISBN: 978-91-8104-439-3
- ISBN: 978-91-8104-440-9
Defence date
25 April 2025
Defence time
09:00
Defence place
Auditorium, building V, Klas Anshelms väg 14, Faculty of Engineering LTH, Lund University, Lund.
Opponent
- Azra Habibovic (Sr Researcher)